SEARCH

SEARCH BY CITATION

Keywords:

  • acoustic method;
  • Anabat;
  • detector;
  • ecoObs-batcorder;
  • heterodyne system;
  • monitoring;
  • Pettersson D240X;
  • Pettersson D500X;
  • Song Meter SM2BAT;
  • transect survey

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

1. The species-rich group of bats fills a wide range of ecological niches and provides ecosystem services like pest control. Bats are known to be sensitive to environmental stressors and could, therefore, be used in assessing ecosystem quality. To use bats as bioindicators, a standardized bat survey method needs to be established as the existing approaches vary in their methodology, and results are, therefore, often not comparable.

2. Generally, there are two different acoustic bat survey methods: the transect walk and the stationary measurement. By conducting transect surveys and simultaneously using several stationary systems, we measured bat activity within a homogeneous habitat and evaluated which method assessed the spatial bat activity patterns with highest precision. Also the survey tool – the detectors themselves – can be grouped into devices with two different methods of triggering the recording of ultrasonic signals: actively by a fieldworker or automatically by a built-in recording control algorithm of the detector. We measured bat activity simultaneously and side by side with both methods for direct comparison.

3. Our results indicate that the transect survey fails to represent the heterogeneous bat activity patterns in a homogeneous landscape. Furthermore, errors occur based on the subjective hearing of the active triggering of the data recording by the human operator.

4. The application of several stationary and automatic sampling systems has the highest potential for standardized acoustic bat surveys. The general use of such an approach would enable us to understand bat activity at landscape scale and could lead to an improvement of bats as bioindicators.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Recently, Jones et al. (2009) argued that bats have great potential as bioindicators. Changes in bat populations or activity were related to climate change, water quality, agricultural intensification, loss and fragmentation of forest habitats, and habitat pollution (Jones et al. 2009).

Most current bat surveys rely on acoustic methods. Contrary to capture methods, telemetry and direct observations, the use of ultrasonic detectors is often the only logistically feasible and cost-effective survey method. However, fundamental concerns regarding the basic methodological designs of many acoustic surveys were expressed as they often fail to address temporal and spatial variation in bat activity patterns (Hayes 2000; Sherwin, Gannon & Haymond 2000; Gannon, Sherwin & Haymond 2003). The assessment of variation in bat activity pattern and the comparability of independent studies are basic requirements for the potential use of bats as bioindicators. Consequently, a standardized bat survey method using a suitable detector system has to be established.

In practice, there are two different methods to survey bat activity: (i) line transects and (ii) the stationary measurement. The transect method is based on the assumption that the bat activity recorded along a transect is in accordance with the activity in the whole habitat of concern (i.e. the habitat is supposed to be homogeneous in regard to bat activity). In contrast, stationary systems are set up at a selected site to reflect the overall activity of the corresponding habitat. To test the transect survey assumption of evenly distributed bat activity within a habitat, we measured bat activity in an agricultural landscape dominated by cereal fields, a habitat assumed to be completely homogeneous, by conducting a transect survey while simultaneously recording with several stationary systems.

In addition to the two different survey methods, there are also two different methods of triggering the recording of ultrasonic signals: (i) actively by a fieldworker using heterodyne or frequency division systems [e.g. Pettersson D200 and D240X (Pettersson Electronic AB, Uppsala, Sweden) or S-25 bat detector (Ultra Sound Advice, London, UK)] or (ii) automatically by the detector devices [e.g. Anabat II (Titley Electronics, Ballina, Australia) or ecoObs-batcorder (ecoObs GmbH, Nürnberg, Germany)]. We measured bat activity simultaneously and side by side with a hand-held Pettersson D240X detector in the heterodyne modus and an automatic ecoObs-batcorder in open landscapes and forests to examine whether differences in data recording occur because of the subjective hearing of the operator of the active-triggering system in comparison with a built-in recording control algorithm of an automatic device.

Material and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

The fieldwork was conducted around Landau (S-Germany). To avoid seasonal differences, bat activity was recorded from June to August (2008–2009). All recordings were obtained during nights with temperatures between 16 and 21 °C, wind speed below 10 km h−1 and without rain.

Addressing Spatial Variability of Bat Activity – Transect Walks vs. Multiple Stationary Measurements

We conducted transect surveys along a 4-km loop trail (Fig. 1) in a homogeneous agricultural landscape consisting mainly of cereal fields. We chose an area free from linear features like woodland edges or hedgerows as bats use them as flight paths (Verboom & Huitema 1997). Bat activity on the loop trail was detected with the heterodyne system of a Pettersson D240X by continuously scanning between 20 and 60 kHz to cover the frequency ranges of expected bat species. High input gain of the detector was selected, and headphones were used to avoid interfering background noise. The first author of this study walked at a constant speed (c. 1 h for the transect) holding the Pettersson detector at a 45° direction relative to the ground and at a height of about 170 cm. Starting 1 h after sunset, the same transect was walked during nine nights from July to August 2009. Simultaneously, three ecoObs-batcorders were installed at a height of 250 cm at equal distances (c.1·3 km) to each other along the loop trail as stationary sampling points (Fig. 1: referred to as sites A–C). The threshold influencing the recording range was set to a fixed sensitivity (full gain at 40 kHz and 96 dB SPL), which resembles a recording radius of about 10 m for most European bat species (Runkel 2008). Calls were determined using sonograms produced with the software bcAnalyze version 1.10 (ecoObs GmbH). Pipistrellus pipistrellus was the only occurring bat species. A pass was defined as a sequence of calls that end five or more seconds before the next sequence begins. The differences in numbers of recorded P. pipistrellus passes were analysed using paired t-tests for each combination of the stationary sampling sites. The numbers of bat passes were normally distributed.

image

Figure 1.  Schematic diagram of the survey design. Transect surveys were performed along a 4-km loop trail (c. 1 h). Simultaneously, bat activity was recorded at three stationary sampling points (A, B, C) by the means of ecoObs-batcorders which were installed at a height of 250 cm at equal distances (c. 1·3 km) to each other along the loop trail.

Download figure to PowerPoint

Comparison of Simultaneous Active and Automatic Triggering of Acoustic Bat Recording

In this approach, bat activity was measured along small forest paths (referred to as forest habitats; n = 12) and homogeneous agricultural areas without any linear elements (referred to as open landscape; n = 27). At each location, bats were detected for 1 h per night simultaneously by an ecoObs-batcorder placed on top of a 170-cm pole (automatic approach) and by a Pettersson D240X held at a height of 170 cm and at a 45° direction relative to the ground (active approach). The sampling points at the forest path were chosen to have at least a radius of 10 m of uncluttered acoustic space around. The ecoObs-batcorder and the Pettersson detector were used as described previously. In the study area, only P. pipistrellus, Nyctalus noctula, Nyctalus leisleri and Eptesicus serotinus were regularly present. Because of the overlapping between acoustic repertoires of the latter three species, it was impossible to assign all call sequences to one of those species with sufficient confidence. Hence, the three species were assigned to the group Nyctalus–Eptesicus. This resulted in two different acoustic groups: P. pipistrellus calling at the upper end of the recordable frequency scale (35–55 kHz) and Nyctalus–Eptesicus calling at lower frequencies (20–35 kHz). The differences in the number of bat passes between both approaches were analysed using paired t-tests for each group of bats and for each habitat type. A bat pass was defined as stated earlier. The numbers of passes were normally distributed.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Addressing Spatial Variability of Bat Activity – Transect Walks vs. Multiple Stationary Measurements

Site A revealed on average 3–5 times the number of P. pipistrellus passes than sites B and C (Table 1A, Fig. 2). Paired t-tests between all three possible combinations of the stationary recording sites in number of detected P. pipistrellus passes showed significant differences in the numbers of passes at the three sites (Table 1B).

Table 1.  Spatial variation of bat activity in a homogeneous landscape. (A) Number of recorded Pipistrellus pipistrellus bat passes at three different stationary sites and on a transect walk in a homogeneous agricultural landscape. Bat activity was measured for 1 h at nine different days between July and August 2009. (B) Statistical differences of P. pipistrellus bat passes between all combinations of the three stationary recording sites. P-values derived from paired t-tests
  n Mean ± SD (minimum–maximum) 
(A)
Site A916·7 ± 0·7 (14–19) 
Site B93·2 ± 0·4 (2–5) 
Site C95·4 ± 0·6 (3–8) 
Transect94·7 ± 2·4 (0–17) 
 d.f. t-value P
(B)
Site A/Site B819·5<0·001
Site A/Site C814·4<0·001
Site B/Site C8−4·80·001
image

Figure 2.  Mean and standard error (bars) of recorded Pipistrellus pipistrellus passes per hour in a homogeneous agriculture landscape in dependence of the number of sample events (sampling in chronological order). Passes were recorded by a Pettersson D240X during a transect survey (inline image) and by three stationary sampling points along the transect by the means of ecoObs-batcorders (Site A: bsl00001; Site B: □; Site C: bsl00123).

Download figure to PowerPoint

Compared with the data of the three stationary devices, detecting of bat activity by walking transects resulted in less precise measures of activity as seen by the large standard errors around the mean value of the number of recorded P. pipistrellus passes (Table 1A; Fig. 2).

Comparison of Simultaneous Active and Automatic Triggering of Acoustic Bat Recording

For passes of Nyctalus–Eptesicus, the simultaneously acquired data sets of both approaches were significantly different, with the active approach always detecting more passes. The difference was more pronounced in the open landscape (t = −6·24, d.f. = 26, P < 0·001; Table 2) than in the forest habitats (t = −2·55, d.f. = 11, P = 0·027; Table 2). On average, the active approach detected 3·6 times more Nyctalus–Eptesicus passes in the open landscape and 1·4 times more Nyctalus–Eptesicus passes in forests in comparison with the automatic ecoObs-batcorder (automatic approach). For P. pipistrellus, the active approach detected significantly more bat passes only in the open landscape (t = −3·22; d.f. = 26, P = 0·003; Table 2), on average 1·1 times more than the ecoObs-batcorder. No statistically significant difference was detected between both methods in the forest habitats (t = −1·15; d.f. = 11, P = 0·275; Table 2).

Table 2. Statistical comparison of the number of bat passes detected by active and automatic triggering of acoustic recording in dependence of the clutterness of the habitat. Bat passes (Nyctalus–Eptesicus and Pipistrellus pipistrellus) were detected side by side and simultaneously by a fieldworker using a Pettersson D240X (active approach) and an ecoObs-batcorder (automatic approach) at several sites in the open landscape and forests for 1 h. The ratio between the means of the active approach to the automatic approach is given. P-values derived from paired t-tests
 Mean ± SD (recorded bat passes)
Active approachAutomatic approachRatiod.f. t-value P
Open landscape
 Nyctalus–Eptesicus 6·9 ± 4·71·9 ± 0·93·626−6,24<0·001
 P. pipistrellus 8·7 ± 7·18·0 ± 6·71·126−3·220·003
Forest
 Nyctalus–Eptesicus 2·2 ± 0·91·6 ± 0·71·411−2·550·027
 P. pipistrellus 6·1 ± 2·35·9 ± 2·51·011−1·150·275

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Sampling Method for Standardized Bat Activity Surveys

Patterns of bat activity may vary within a habitat in response to a variety of biotic (e.g. abundance of prey insects) and abiotic (e.g. landscape structure) factors. Our results demonstrated significantly different bat activity patterns even in a homogeneous agricultural landscape. We assume that spatial heterogeneity of bat activity is much more pronounced in habitats with higher structural heterogeneity such as forests (within a forest there are clearings, young undergrowth, mature trees). Measurements from transect surveys give information about bat activity on the landscape scale, but fail to account for spatial variation and, therefore, may miss small but important foraging areas that could result in inappropriate management recommendations. Moreover, the transect method cannot assess spatial and temporal activity at a given site and, therefore, may miss vital periods of bats at certain parts of the transect. The stationary and simultaneous measurement at a number of detecting sites within a habitat accounts for this spatial and temporal variation of bat activity and is, therefore, the recommended approach for a standardized bat activity survey.

Sampling Device for Standardized Bat Activity Surveys

In acoustic surveys, bat activity cannot be compared across species (Jones, Vaughan & Parsons 2000) because of species-specific differences in frequency ranges and intensity of echolocation that are subject to different levels of frequency-dependent atmospheric attenuation (Lawrence & Simmons 1982). For example, the longer calls, the lower amplitude and the less rapid attenuation of the lower frequency calls of Nyctalus–Eptesicus result in an enlarged detection range of that group compared with P. pipistrellus. Nevertheless, the comparison of activity is possible for each species independently as long as the detection capability is constant across all habitats. We compared simultaneously performed active and automatic triggering of the detecting of two acoustically different bat groups in two habitats that differ in the amount of acoustic clutter and demonstrated different detection ratios of both groups depending on the habitat. Systems with automatic triggering of the acoustic detection such as the ecoObs-batcorder have a definite and standardized detection amplitude threshold (Runkel 2008), whereas the human operator, by using a heterodyne system, can detect ‘expected’ signals even within the background noise. Thus, compared with the forest where higher level of clutter leads to scattering, spreading, absorption and reflection of the echolocation calls (Griffin 1971; Parsons 1996), in the open landscape, low-frequency calls arriving from a greater distance can be detected by the fieldworker (active approach) quite well. This expanded detection range of the active approach results in an overestimation of the activity of bat species using low frequency calls (e.g. Nyctalus-Eptesicus) in uncluttered habitats such as the open landscape. This bias of the active approach violates the assumption that the amount of bat calls of a given species detected at a site reflects the intensity of habitat use, and comparisons between different habitats are, therefore, misleading. In addition, the active-triggering approach depends on the skills, the subjective hearing ability, concentration or tiredness of the operator in the field leading to further potential biases.

Active triggering is labour-intensive, as it requires one person per sampling site and detector, whereas several automatic recording units can be handled by a single person alone. In the active approach, it is, therefore, common practice that bat activity is only measured during a certain nocturnal time span (e.g. Kusch et al. 2004). But temporal activity patterns of bats may considerably vary on a nightly basis in response to a variety of factors including abundance of prey insects and weather conditions (Hayes 1997; O’Donnell 2000). Moreover, the output signals of heterodyne systems as used in active-triggering approaches contain no information on the duration and frequency of the original input signal and are, therefore, unsuitable for further bioacoustic analysis.

In contrast, automatic triggering seems to be a promising tool for an unbiased and comparable assessment of bat activity (see Table 3 for a comparison with the active approach). However, the range of automatically triggered recording devices available differs in the extent to which they are suitable for a standardized bat activity survey. Simple automatic systems for recording of bat activity can be assembled by connecting a voice-activated tape recorder to a heterodyne detector (e.g. O’Donnell 2000) or a time-expansion detector, respectively. In the heterodyne system, a frequency dial has to be set, limiting this approach to the detection of bat calls within the selected narrow frequency range (Parsons & Szewczak 2009). During the output phase of the time-expansion system, it is not possible to record further bat calls (Parsons & Szewczak 2009). Thus, both types of self-assembled automatic systems are unsuitable for a standardized bat survey method.

Table 3.  Overview of the different commercially available approaches for the staionary measurement of bat activity in regard to their labour intensity, quality of the acoustic data, comparability of results and purchase prize. Self-assembled automatic systems are not considered
  1. *If used with exposed microphone and calibrated (only the ecoObs-batcorder fulfils that in its standard version).

Triggering of the detectionActiveAutomaticAutomatic
Transforming of the ultrasound prior to recordingHeterodyningZero-crossingDirect recording
ExamplesPettersson D200Anabat IIecoObs-batcorder
S-25 detector Pettersson D500X
  SM2BAT
Labour intensityHighLowLow
Quality of dataLowLimitedHigh
Comparability of the resultsNot possible (subjective biases)Not possible (directionality)Possible*
Purchase price c. 300 € c. 2000 €800–3000 €

The remaining automatic detector systems can be divided into zero-crossings period meters (Anabat II) and recently developed detectors systems with high-speed analogue to digital converters that can directly record ultrasound (real time recorder) such as the ecoObs-batcorder, the Pettersson D500X (Pettersson Electronic AB) and the Song Meter SM2BAT (Wildlife Acoustics; Concord, USA). We discuss the suitability of these detector systems to meet the demands of a standardized stationary bat survey in the following (see Table 3 for an overview).

Quality of the recording data

Analysis of zero-crossing period meters shows only information about the strongest harmonic in any signal (Fenton et al. 2001), but fine spectrotemporal details are missing. This reduction can complicate species determination within some genera (Fenton 2000; Parsons, Boonman & Obrist 2000; Parsons & Szewczak 2009). Ultrasound can be recorded directly with real-time recorders, and untransformed and uncompressed data files such as WAV files are produced. The preservation of all characteristics of the original signal allows detailed bioacoustical descriptions of bat calls (Jones, Vaughan & Parsons 2000; Parsons & Szewczak 2009).

By considering the vast amount of call sequences one can get with several autonomous recording units, the manual measurement of call characteristics becomes a bottleneck. To cope with this problem, multivariate identification software can be applied (Jones, Vaughan & Parsons 2000; Parsons & Szewczak 2009). These identification systems rely on the high-quality data of real-time recording (Parsons & Szewczak 2009). Apart from saving time and removing any subjectivity from species identification (Jones, Vaughan & Parsons 2000), a further advantage of identification systems is that they can also be used by nonspecialists in acoustics.

Recent studies use the feeding buzz rate of bats as an indicator of the value of particular areas as foraging habitats (e.g. Vaughan, Jones & Harris 1997). A feeding or final buzz is in general a sequence of calls becoming shorter in duration and broader in bandwidth and is emitted by aerial feeding bats during prey-capture attempts (Griffin, Webster & Michael 1960). Feeding buzzes are especially faint and difficult to record with the Anabat II (Scott et al. 2010) compared with detectors that use direct recording as only strong portions of any call will activate the zero-crossing period meter (Fenton et al. 2001).

Comparability of results

To compare results obtained by more than one recording unit, detection fields of identical size and shape are required.

Individual units of the same detector system can vary extensively in the size of their detection fields (Waters & Walsh 1994; Fenton 2000; Larson & Hayes 2000). The units can be calibrated against each other by using an external ultrasonic sound (e.g. Larson & Hayes 2000 for the Anabat II system), but fine-scale adjustments require a signal generator in a controlled laboratory environment. The ecoObs-batcorder is so far the only commercially available detector system that is calibrated in a standardized way (Runkel 2008) and, therefore, allowing direct comparisons of independent studies.

The shape of the detection field depends on the directionality of the microphone. The directionality itself is a function of the microphone type and its position. Condenser microphones and recessed microphone positions increase directionality, while miniature electret microphones and exposed microphone positions allow greater omni-directionality (Pye 1993; Waters & Walsh 1994; Runkel 2008). Because of the condenser microphone and its fixed position on the detector surface, the Anabat II shows high directionality, resulting in a detection field being longer than wide (Larson & Hayes 2000). This can lead to differences in the number of recordings at the same site because of the orientation of the detector and may result in biases between sites (Weller & Zabel 2002). The Pettersson D500X, SM2BAT and ecoObs-batcorder have omni-directional miniature electret microphones. While there are options to use an external microphone extended on a cable in the Pettersson D500X and SM2BAT system, the ecoObs-batcorder already has an external microphone in an exposed position in its standard design.

Conclusion

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Considering our results, the use of several randomly selected stationary sampling points within a habitat using calibrated and automatically triggered real-time recording devices (e.g. ecoObs-batcorder) has the highest potential for standardized acoustic bat surveys. The proposed approach produces unbiased and comparable data sets on the relative activity of bats. In addition, the use of stationary and automatic recording systems is less labour-intensive and time-consuming and even feasible for nonbat-specialists and, therefore, represents a cost-effective survey method.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

We especially thank Volker Runkel who had a lot of technical input on the manuscript. We are grateful to M. Brock Fenton and an anonymous reviewer for useful comments and suggestions that helped to improve the manuscript. We also thank Björn Siemers for a critical review and valuable comments, John K. Tucker for improvements regarding the language and Klaus Swarowsky for help with the figures.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References
  • Fenton, M.B. (2000) Choosing the ‘correct’ bat detector. Acta Chiropterologica, 2, 215224.
  • Fenton, M.B., Bouchard, S., Vonhof, M.J. & Zigouris, J. (2001) Time-expansion and zero-crossing period meter systems present significantly different views of echolocation calls of bats. Journal of Mammalogy, 82, 721727.
  • Gannon, W.L., Sherwin, R.E. & Haymond, S. (2003) On the importance of articulating assumptions when conducting acoustic studies of habitat use of bats. Wildlife Society Bulletin, 31, 4561.
  • Griffin, D.R. (1971) The importance of atmospheric attenuation for the echolocation of bats. Animal Behaviour, 19, 5561.
  • Griffin, D.R., Webster, F.A. & Michael, C.R. (1960) The echolocation of flying insects by bats. Animal Behaviour, 8, 141154.
  • Hayes, J.P. (1997) Temporal variation in activity of bats and the design of echolocation-monitoring studies. Journal of Mammalogy, 78, 514524.
  • Hayes, J.P. (2000) Assumptions and practical considerations in the design and interpretation of echolocation-monitoring studies. Acta Chiropterologica, 2, 225236.
  • Jones, G., Vaughan, N. & Parsons, S. (2000) Acoustic identification of bats from directly sampled and time-expanded recordings of vocalizations. Acta Chiropterologica, 2, 155170.
  • Jones, G., Jacobs, D.S., Kunz, T.H., Willig, M.R. & Racey, P.A. (2009) Carpe noctem: the importance of bats as bioindicators. Endangered Species Research, 8, 93115.
  • Kusch, J., Weber, C., Idelberger, S. & Koob, T. (2004) Foraging habitat preferences of bats in relation to food supply and spatial vegetation structures in a western European low mountain range forest. Folia Zoologia, 53, 113128.
  • Larson, D.J. & Hayes, J.P. (2000) Variability in sensitivity of Anabat II bat detectors and a method of calibration. Acta Chiropterologica, 2, 209213.
  • Lawrence, B.D. & Simmons, J.A. (1982) Measurement of atmospheric attenuation in ultrasonic frequencies and the significance for echolocation by bats. Journal of the Acoustical Society of America, 71, 585590.
  • O’Donnell, C.F.J. (2000) Influence of season, habitat, temperature, and invertebrate availability on activity of the New Zealand long-tailed bat (Chalinolobus tuberculatus). New Zealand Journal of Zoology, 27, 207221.
  • Parsons, S. (1996) A comparison of the performance of a brand of broad-band and several brands of narrow-band bat detectors in two different habitat types. Bioacoustics, 7, 3343.
  • Parsons, S., Boonman, A.M. & Obrist, M.K. (2000) Advantages and disadvantages of techniques for transforming and analyzing chiropteran echolocation calls. Journal of Mammalogy, 81, 927938.
  • Parsons, S. & Szewczak, J.M. (2009) Detecting, recording, and analyzing the vocalizations of bats. Ecological and Behavioral Methods for the Study of Bats, 2nd edn (eds T.H. Kunz & S. Parsons), pp. 91111. Johns Hopkins University Press, Baltimore.
  • Pye, J.D. (1993) Is fidelity futile? The true signal is illusory, especially with ultrasound. Bioacoustics, 4, 271286.
  • Runkel, V. (2008) Mikrohabitatnutzung syntoper Waldfledermäuse. PhD thesis, Universität Erlangen-Nürnberg.
  • Scott, S.J., McLaren, G., Jones, G. & Harris, S. (2010) The impact of riparian habitat quality on the foraging and activity of pipistrelle bats (Pipistrellus spp.). Journal of Zoology, 280, 371378.
  • Sherwin, R.E., Gannon, W.L. & Haymond, S. (2000) The efficacy of acoustic techniques to infer differential use of habitat by bats. Acta Chiropterologica, 2, 145153.
  • Vaughan, N., Jones, G. & Harris, S. (1997) Habitat use by bats (Chiroptera) by means of a broad-band acoustic method. Journal of Applied Ecology, 34, 716730.
  • Verboom, B. & Huitema, H. (1997) The importance of linear landscape elements for the pipistrelle Pipistrellus pipistrellus and the serotine bat Eptesicus serotinus. Landscape Ecology, 12, 117125.
  • Waters, D.A. & Walsh, A.L. (1994) The influence of bat detector brand on the quantitative estimation of bat activity. Bioacoustics, 5, 205221.
  • Weller, T.J. & Zabel, C.J. (2002) Variation in bat detections due to detector orientation in a forest. Wildlife Society Bulletin, 30, 922939.