OUP user menu

Variation of Echolocation Calls of Pteronotus quadridens (Chiroptera: Mormoopidae) in Cuba

Silvio Macías, Emanuel C. Mora
DOI: http://dx.doi.org/10.1644/BWG-019 1428-1436 First published online: 21 November 2003


Echolocation calls were recorded from Pteronotus quadridens flying in the field and in an enclosed space. In the field, search calls contained 1 or 2 harmonics. Patterns of call design show a segment of quasi-constant frequency (QCF2nd-harmonic at 81–84 kHz), followed by a downward frequency-modulated (FM) component. The 2nd harmonic was always more intense than the 1st. Search, approach, and terminal phases of calls were described during hunting sequences of P. quadridens. The transition between call phases was characterized by monotonic variations in some acoustic parameters, including a decrease in call duration and an increase in repetition rate, bandwidth, and slope of the FM component. We also analyzed calls emitted by bats flying in confined spaces that consistently contained 3 harmonics, of which the 2nd harmonic contained the greatest energy. The values of call duration were shorter and bandwidth was higher than values characterizing calls emitted during the search phase in the field.

Key words
  • bats
  • call variation
  • echolocation
  • Pteronotus quadridens

The sooty mustached bat (Pteronotus quadridens) is a small (forearm length 35–41 mm, weight 3–6 g) mormoopid bat that is well distributed throughout the Greater Antilles (Silva 1979). In Cuba, it is one of the most abundant cavernicolous species. Usually P. quadridens begins foraging about 10 min before sunset and continues throughout the evening, eating insects captured exclusively in flight (Mora et al. 2002; Silva 1979). However, little is known about its echolocation behavior. Schnitzler et al. (in litt.) have briefly discussed echo-location calls emitted by this species while foraging in a flight cage. Their recordings indicate that P. quadridens emits short calls (3.1 ms) that consist of 2 components. First is a constant-frequency component found between 77 and 83 kHz, then a frequency-modulated (FM) component with a band width of 15–22 kHz. Kössl et al. (1999a) described 2 spectrograms of echolocation calls recorded from a caged P. quadridens from Jamaica with a constant-frequency component at 80 kHz.

Insectivorous bats in the family Mormoopidae exhibit the lowest levels of variability in echolocation call design among bat species thus far studied (Fenton 1994; Ibáñez et al. 1999; Vater 1987). Call structure is similar even among different species of this family. Search calls emitted by most mormoopid bats are characterized by a constant-frequency or quasi–constant frequency (QCF) component at the beginning of the signal, followed by an FM component (Ibáñez et al. 1999; O'Farrell and Miller 1997; Schnitzler et al. 1987; Simmons et al. 1979). This type of signal is used by bats that search for insects in areas of high structural clutter, where they must detect and identify insects while avoiding collisions with obstacles (Schnitzler and Kalko 1998; Simmons and Stein 1980).

To our knowledge, no previous study has evaluated possible changes in echolocation calls of P. quadridens. In this paper, we describe and compare calls emitted by free-flying P. quadridens searching for prey and flying in an enclosed space.

Materials and Methods

Echolocation calls from P. quadridens during foraging activity were recorded on 2 nights. Six bats were recorded flying at heights of up to 7 m above ground level, ∼500 m from their roost cave, Cueva del Indio. The cave, containing thousands of individuals, is located 40 km SE of Havana, Cuba. P. quadridens hunted in a small area that could be visually tracked for several minutes while recording under ambient light conditions (crepuscular and moonlight).

After recording echolocation calls of free-flying bats, we captured and identified individuals (Silva 1979). Four bats were captured with hand nets in Cueva del Indio cave, allowed to fly individually in a bat-free section of the cave (4 by 3 by 3 m), and recorded. All recordings were made while bats were flying toward the microphone.

Echolocation calls were recorded with an ultrasound detector (U30 Ultrasound Advice, London, United Kingdom; sensitivity flat, ±2 dB, between 20 and 200 kHz). During recording sessions, heterodyne output was tuned to 82 kHz. The detector's high-frequency output was fed to the analog–digital input port of the digital signal processing board (model PCM-DAS 16S/330, Plug-In Electronic, Eichenau, Germany) of a laptop computer. The board was controlled with commercial software (BatSound 2.1, Pettersson Elektronik, Sweden). Sampling frequency was set at 312 kHz, and storage time was set at 5 s in a cyclical buffer.

Echolocation calls were displayed simultaneously as spectrograms and temporal digitized recordings (oscillograms) with BatSound 2.1. Spectrograms were made of consecutive fast Fourier transforms (FFTs) with a 99% overlap. Usually, a 512-point FFT was chosen in order to get a frequency resolution of 610 Hz. Time was measured on the oscilloscope screen manually. On oscillograms and spectrograms, time resolution was 0.1 ms. To obtain power spectra, FFTs were calculated with 256–1,024 data points.

The parameters measured for each call that registered a maximal intensity >20 dB above noise level were duration, time between start and end of a call (ms) on the oscillogram; peak frequency (kHz), corresponding to maximal intensity in the power spectrum; minimal and maximal frequencies, measured 20 dB below maximal intensity in the power spectrum; bandwidth, calculated as the difference between maximal and minimal frequencies; Q10-dB, calculated as peak frequency divided by bandwidth measured 10 dB below maximal intensity; QCF; and slope of the frequency modulation, calculated as the difference (kHz) between initial and final frequencies of the FM portion divided by duration of the FM portion. QCF and the slope of the frequency modulation were measured from spectrograms (Fig. 1). In each call sequence, we measured intercall interval from the beginning of a call to the start of the next call. Because we were unable to exclude the possibility that our sample included multiple recordings from the same individual (pseudoreplication), parameters of each call sequence of a bat were averaged and treated as a single measurement before further analysis (W. L. Gannon et al., in litt). Call sequences with <5 calls, with calls of insufficient signal to noise ratio (peak intensity <20 dB above noise level measured in power spectrum), or both were not analyzed. Call sequences recorded from free-flying bats can include search, approach, and terminal phase calls, whereas those recorded from animals flying in an enclosed space are only composed of orientation calls.

Fig. 1

Spectrogram of a representative search echolocation call with defined measurement points: quasi–constant frequency (QCF), initial frequency (IF), final frequency (FF), duration of the QCF component (dQCF), and duration of the frequency-modulated component (dFM).

All acoustic parameters are presented as mean ± 1 SD. The ranges of values given for call duration and bandwidth and slope of the frequency modulation correspond to minimal and maximal means calculated in each call sequence. We used Student's t-test to compare 2 mean values. All analyses were made at a significance level of α = 0.05.


The description of echolocation calls emitted by P. quadridens include 10 call sequences (116 calls) of its foraging behavior in 6 free-flying animals and 4 sequences (22 calls) of 4 bats flying in an enclosed space. Two sequences recorded from free-flying animals include approach and terminal phases.

The sequence of echolocation calls emitted by P. quadridens during an insect capture in free-flying bats (Fig. 2) contains all 3 phases (search, approach, and terminal) described by Griffin (1958).

Fig. 2

Echolocation calls emitted by Pteronotus quadridens during a hunting sequence by a free-flying bat, showing the search (s), approach (a), and terminal (tp) phase or buzz. Vertical dashed lines indicate phase transitions. A) Oscillogram and B) spectrogram.

During the search phase, a shallow or QCF-modulated segment characterizes the beginning of each harmonic (1st or 2nd) contained in each echolocation call; this QCF component is followed by an FM component (Fig. 3A). The definition of the QCF component at the beginning of a call follows the classification of Kalko and Schnitzler (1993) as signal elements with a bandwidth of ≤4 kHz and duration >1 ms. In each search call, peak intensity of the 2nd harmonic was >20 dB higher than that of the 1st harmonic (Fig. 3A). For that reason, we only measured parameters that characterize this harmonic in describing calls (Table 1). In search calls, peak frequency was 82.41 ± 0.45 kHz (QCF component). Duration of calls varied between 3.9 and 4.4 ms, and bandwidth was between 15.1 and 16.3 kHz. The slope of frequency modulation ranged from 5.1 to 7.2 kHz/ms.

Fig. 3

Typical echolocation calls emitted by Pteronotus quadridens during A) search, B) approach, and C) terminal phase while hunting for insects. Spectrograms (left) and power spectra (right).

View this table:
Table 1

Comparison between pulses emitted by Pteronotus quadridens during the search phase of its foraging activity (n = 10 passes) and orientation calls emitted while flying in an enclosed space (n = 4 passes). The results of a Student's t-test are shown (FM, frequency modulated; QCF, quasi–constant frequency; ns, not significant).

Search callsOrientation calls
Duration (ms)<0.01
FM slope (kHz/ms)6.120.648.800.48<0.01
QCF duration (ms)1.590.241.690.020.42 (ns)
QCF value (kHz)83.050.6685.720.27<0.01
Peak frequency (kHz)82.410.4585.190.51<0.01
Minimum frequency (kHz)67.70.6165.350.92<0.01
Maximum frequency (kHz)83.500.4785.900.25<0.01
Bandwidth (kHz)15.790.4420.540.790.01
Interpulse Interval (ms)53.333.6871.2718.100.02

The approach phase was defined by calls with a 3rd harmonic of higher frequency (between 123 and 105 kHz; Figs. 2B and 3B) and a tendency to increase call bandwidth (15.1–20.1 kHz). This implies a bandwidth 4–5 kHz higher than calls emitted in search phase calls. Also, there is an increase in the slope of frequency modulation (6.1–9.6 kHz/ms) and in relative intensities of 1st and 3rd harmonics (Fig. 3B). Duration of calls in the approach phase is reduced, reaching values near 3 ms.

The terminal phase starts with a sudden increase in call repetition rate (see Fig. 2A) and a decrease in call duration. Interpulse interval decreases to 5 ms, whereas call duration decreases to 1.2 ms. Calls emitted in the terminal phase are FM signals (Fig. 3C) and do not have a QCF component present in search and approach calls (Figs. 3A and 3B). Variations in temporal and spectral parameters of calls during the 3 phases of an insect pursuit sequence are shown in Fig. 4. Duration of calls during the search phase varies between 4 and 5 ms and decreases in the approach phase. At the beginning of the terminal phase, there is an abrupt decrease in call duration. Repetition rate is rather constant during the search phase and increases abruptly and monotonously during the terminal phase. These 2 parameters behave inversely (Fig. 4A). On the other hand, slope of the frequency modulation and bandwidth show a similar behavior, both increasing monotonically during the hunting sequence.

Fig. 4

Temporal variations of echolocation call parameters in a hunting sequence in Pteronotus quadridens. Vertical dashed lines show limits of search (s), approach (a), and terminal (tp) phases.

Call duration is highly correlated with interpulse interval, (correlation coefficient r = 0.92, P < 0.05, n = 104 calls). Low values for interpulse interval (∼5 ms) correspond to lowest call duration values (between 1 and 2 ms), both appearing in the terminal phase. Highest values of both variables belong to calls of the search phase (intercall interval 59 ms, call duration 4 ms). Bandwidth is significantly correlated with call duration (r = −0.70, P < 0.05, n = 108 calls); an increase in duration corresponds with a decrease in bandwidth.

When P. quadridens flies in an enclosed space, it produces calls with a similar design (QCF-FM) to those emitted during the search phase of its foraging behavior in the wild (Fig. 5). However, these signals always contain 3 harmonics with <20 dB of intensity difference among them. Each harmonic starts with a QCF component followed by an FM component, and peak frequency is localized in the 2nd harmonic. Peak intensity in this case was localized also in the QCF component of the call (85.19 ± 0.51 kHz).

Fig. 5

Typical echolocation call emitted by Pteronotus quadridens during flight in a confined space. Spectrogram (left) and power spectrum (right).

Recordings of echolocation calls emitted by P. quadridens in different flight conditions suggested variations in structure of the calls. To verify this, we used a Student's t-test to compare calls emitted in the wild with those produced in an enclosed space (Table 1). Duration of calls in the enclosed space of the cave and Q10-dB showed lower values, whereas slope of the frequency modulation and bandwidth had higher mean values. Thus, P. quadridens increments frequency content of its calls when flying in an enclosed space.


The design of echolocation calls used by P. quadridens living in Cueva del Indio, Cuba, during foraging activity is similar to that described in Jamaica by Schnitzler et al. (1987) and Kössl et al. (1999a). Call design is also similar to that characterized in other species of Pteronotus living in Jamaica and elsewhere (Ibáñez et al. 1999, 2000; O'Farrell and Miller 1997; Schnitzler et al. 1987; Schnitzler and Kalko 1998; Simmons et al. 1979). Call design consists of a component of QCF followed by an FM component.

The QCF component could be associated with detection of targets and evaluation of prey movements by Doppler shift. A similar function has been attributed to QCF search calls of insectivorous bats that hunt in open spaces (e.g., Tadarida brasiliensisSimmons et al. 1979; Diclidurus albus, PeropteryxKalko 1995b; Myotis negricansSiemers et al. 2001; Molossus molossusKössl et al. 1999b; E. C. Mora et al., in litt.).

Pteronotus quadridens feeds exclusively on flying insects (e.g., Coleoptera, Diptera, Lepidoptera, and Orthoptera) captured in environments with background clutter (Silva 1979). The QCF component of calls might improve the chances of long-range detection and identification of prey by bats (Kalko and Schnitzler 1993; Schnitzler et al. 1987; Surlykke et al. 1993). Moreover, detection of prey is improved when a call contacts an insect at the exact instant when its wings are perpendicular to the impinging sound wave. At that moment, an acoustical glint is produced that can be perceived by the bat. Glint detection increases the chance of detecting flying insects (Schnitzler et al. 1987). Narrowband signals are well suited for detecting insect echoes because they concentrate sound energy in a small frequency band (Schnitzler et al. 1987). It has been demonstrated that bats that use long constant-frequency signals, like horseshoe bats and P. parnelii, not only detect potential prey by their wingbeat but also can perceive and evaluate the wing beat rate of their prey (Emde and Schnitzler 1986). Thus, narrowband call elements are adapted for medium-range detection of insects flying in front of background clutter and for delivering some flutter information.

On the other hand, FM components of search phase calls probably provide information about objects included in background clutter, such as vegetation, as well as about prey, once detected (Simmons et al. 1979; Simmons and Stein 1980). FM components might also deliver spectral cues that describe the nature of the reflecting target (Schmidt 1988). Combining QCF-FM components could be an obstacle-monitoring strategy that combines the advantages of a single multipurpose echolocation signal (Simmons and Stein 1980). Some bats, such as those of the genus Pipistrellus and of Eptesicus fuscus (Vespertilionidae), emit an FM component that precedes the QCF component (FM-QCF—Schnitzler and Kalko 1998). This kind of strategy differs from that used by many vespertilionid bats that forage in clutter, such as M. evotis, which are effectively foraging for insects against clutter using FM calls exclusively by continuously sweeping from 90 to 30 kHz (Faure and Barclay 1994). Like QCF-FM calls, short broadband signals used by species of vespertilionid bats might help localize both prey and nearby obstacles for directed approach or avoidance, respectively (Siemers and Schnitzler 2000).

In the approach and terminal phases of its hunting sequences, P. quadridens reduces or eliminates the QCF component of its calls, which in the search phase is likely to facilitate detection of prey and evaluation of prey movements (Kalko and Schnitzler 1993). As shown for Pipistrellus pipistrellus, P. kuhli, and P. nathusii (Kalko and Schnitzler 1993), P. quadridens should detect prey only when the echo that it receives from the prey does not overlap with the outgoing signal or with other echoes produced by background clutter such as vegetation.

Calls of the terminal phase are best suited for determining distance and prey position because of their short, wideband, steep FM signals that often include a 3rd harmonic (Fenton and Bell 1979; Griffin 1958; Kalko 1995a; Kalko and Schnitzler 1989, 1993; Schnitzler et al. 1987; Simmons and Stein 1980). In P. quadridens, 3rd harmonics increase call bandwidth. This is a different strategy than that used by vespertilionid bats, which increase bandwidth by increasing the FM component (Siemers et al. 2001; Surlykke et al. 1993; Surlykke and Moss 2000). However, other species, such as Saccopteryx bilineata, S. leptura, Cormura brevirostris (Emballonuridae), and Craseonycteris thonglongyai (Craseonycteridae) maintain short constant-frequency or QCF components in their signals during the approach and terminal phases (Barclay 1983; Kalko 1995b; Surlykke et al. 1993). As mentioned before, P. quadridens would only detect prey while emitting calls with a duration and repetition rate that match the overlap-free window. Changes in call duration and repetition rate that occur in approach and terminal phases could be a strategy to avoid self-deafening. Also, a high repetition rate enhances information content per unit time, which enables the bat to prepare for the catch.

Schnitzler et al. (1987) described search calls of P. quadridens in Jamaica with a duration of 3.1 ms, a QCF segment in the 2nd harmonic at frequencies between 77 and 83 kHz, and an FM component with bandwidth between 15 and 22 kHz. In our recordings in free-flying bats, search calls had durations of 4.2 ms, a QCF component between 81.9 and 84.2 kHz, and a bandwidth between 15.1 and 16.3 kHz. We think that differences in call duration and bandwidth between these 2 studies are due to differences in flight conditions. Schnitzler et al. (1987) reported values for calls during hunting sequences recorded in flight cages, whereas we recorded calls during foraging by free-flying bats. We obtained similar values to Schnitzler et al. (1987) when Cuban P. quadridens were flying and recorded in an enclosed space (duration 3.2 ms, bandwidth 20.5 kHz). Thus, we suggest that rather than latitude, use of an enclosure during recording provided calls with shorter duration and greater bandwidth than echo-location calls normally found from free-flying bats. Nevertheless, QCF values in calls emitted from captive animals in enclosed spaces in Cuba (85.7 kHz) was higher than the upper limit described for search calls of bats in Jamaica. Combining QCF values recorded from Cuban P. quadridens that were both free-flying and flying in an enclosure, we report higher values than for this species from Jamaica, which agrees with the inverse relationship between body size and frequency (Jones 1999). Animals from Jamaica are larger than those from Cuba (length of forearm 38.0 ± 0.34 and 37.3 ± 0.26 mm, respectively; P < 0.001, t = 7.8—Smith 1972) and produce lower QCF values. However, size differences should be explored further relative to these insular species (Guillén et al. 2000).

Echolocation calls of P. quadridens are different from calls of other mormoopids, including P. parnelii and Mormoops blainvillei (Kössl et al. 1999a; E. C. Mora and S. Macias, in litt.; Schnitzler et al. 1987; Schnitzler and Kalko 1998), but are similar to P. macleayi (Kössl et al. 1999a). However, P. quadridens has a higher frequency range (67.7–83.5 kHz) of the 2nd harmonic than P. macleayi (55–71 kHz).

Although call structure of P. quadridens is unique, variations were found. QCF was higher for bats in an enclosed space than for free-flying bats (85.7 versus 83.0 kHz, respectively), bandwidth increased (20.5 versus 15.8 kHz, respectively), and call duration decreased (3.2 versus 4.2 ms, respectively). Bats with echolocation systems that rely on narrowband frequency processing (such as paleotropical hipposiderids and rhinolophids or eotropical mormoopids) typically show limited behavioral plasticity in their calls (Fenton 1994; Vater 1987). Nonetheless, variation in parameters of orientation calls used by P. quadridens when flying in enclosed spaces, such as caves, suggests that call design in this species is important for foraging and navigation.


The authors express their gratitude to C. Ibañez and J. Juste, Sevilla, Spain, who kindly helped us with literature pertaining to this study and for an earlier revision of the manuscript. Special thanks to F. Coro and M. Kössl for their valuable comments on this manuscript. The equipment used in this work was provided by a grant from the VW Foundation project (I/77306). Our gratitude goes as well to 2 anonymous reviewers who made very important contributions to earlier versions of the manuscript.


  • Associate Editor was William L. Gannon.

Literature Cited

View Abstract