Bats of Southern and Central Africa. Ara Monadjem

1.4±0.4 0.8±0.07 0.3±0 Myotis tricolor 3 58.1±5.3 38.1±3.8 50.7±1.1 54.3±1.8 1.7±1.3 0.5±0.2 0.2±0.3 Neoromicia capensis 2 52.7±11.9 39.3±0.7 41.7±0.1 44.9±4.4 1.3±0.1 0.8±0.4 0.2±0 Neoromicia nana 3 86.0±10.4 67.9±4.3 75.3±7.8 81.4±14.6 1.3±1.2 0.9±0.8 0.2±0.2 Neoromicia zuluensis 4 62.8±5.4 48.8±3.7 50.4±3.7 54.8±5.9 2.7±1.2 2.4±1.5 1.0±0.6 Nycticeinops schlieffeni 8 50.6±4.7 39.4±3.1 41.1±1.9 43.8±2.2 2.2±0.9 2.0±0.9 0.8±0.4 Pipistrellus hesperidus 10 65.4±5.7 46.9±2 50.4±1.9 54.8±2.9 2.0±0.7 2.5±1.4 1.0±0.5 Scotophilus dinganii 11 44.2±6.6 33.6±2.5 34.0±2.8 36.6±2.9 3.0±1.1 2.8±1.1 1.2±0.6 Scotophilus viridis 4 57.5±5.6 41.1±4.2 42.9±4.8 46.3±4.6 3.7±2.7 2.3±1.8 0.7±0.4

      Time expansion detectors digitise bat calls at a high sampling rate and replay them at a lower sampling rate afterwards. Typically, the sampling rate ratios vary from 1:10 to 1:32. This is the equivalent of recording sound on a high-speed tape recorder and then playing it back at a slower speed – time is effectively ‘expanded’ (slowed down) by a set factor (e.g. 10). Real-time full-spectrum detectors record the full frequency range up to a limit determined by the sampling rate used (where maximum recordable frequency is half the recording sampling rate, i.e. if the detector’s sampling rate is, e.g., 256 kHz, it will record sounds up to 128 kHz). The main disadvantage of time expansion and real-time full-spectrum bat detectors is that they are usually expensive. The main advantage of these detectors is that all information of bat calls – including amplitude, frequency and harmonic structure – is preserved, making it ideal for detailed analyses of call characteristics. Echolocation data used to generate the spectrograms in this book were recorded with time-expansion and real-time full-spectrum bat detectors: the Pettersson D980 detector (Pettersson Elektronik AB, Uppsala, Sweden) and the Avisoft Ultrasoundgate 416 and 116 detectors fitted with Ultrasoundgate CM16 microphones (Avisoft Bioacoustics, Berlin, Germany).

      Note that some modern bat detectors are capable of recording in multiple modes.

      Sampling bat diversity using bat detectors. Bat detectors are increasingly being used to survey bats for scientific research (e.g. Schoeman and Waddington 2011, Taylor et al. 2013a, Schoeman 2016, Mtsetfwa et al. 2018, McCleery et al. 2018), and acoustic monitoring has been recommended as the principal way of surveying bats for environmental impact assessments in South Africa (Sowler and Stoffberg 2012). To use bat detectors effectively, it is imperative to develop ‘local call libraries’ (i.e. recordings of echolocation calls from a particular locality) because these calls may vary geographically (e.g. Stoffberg et al. 2012). Without a call library, the chances of correctly identifying bats by their echolocation calls can be difficult, particularly in species-rich and poorly sampled regions. Examples of call libraries for southern Africa include those by Schoeman (2006), Taylor et al. (2013b) and Monadjem et al. (2017). In addition, bat calls, like all waves, attenuate with distance and this is proportional to the frequency of the call; the further away the bat detector, the greater the chance of missing high-frequency calls. Consequently, bat communities sampled using only bat detectors will probably underrepresent bats with high-frequency calls. A correction factor can improve the probability of acoustic surveys representing the actual bat community (see Monadjem et al. (2017) for a correction factor for southern African bats).

SPECTROGRAMS AND ECHOLOCATION

      In this book, three criteria were applied to graphically represent the echolocation call of bat species on spectrograms using BatSound Pro software (version 3.20; Pettersson Elektronik AB, Uppsala, Sweden). First, only signals with a high signal-to-noise ratio were considered, that is, the signal from the bat was at least three times stronger than the background noise as displayed on the time-amplitude window. Second, only signals that were not saturated were analysed (Fenton et al. 2001). Finally, for most low duty-cycle bats only search-phase signals that were recorded at least three seconds after releasing the bat were considered (O’Farrell and Gannon 1999). Echolocation calls of high duty-cycle bats were recorded with the bat in the hand to eliminate any possible Doppler-shift or compensation for it by the bat when in flight (Heller and von Helversen 1989).

      Duration, bandwidth and peak echolocation frequency are reported as indicators of sensory ability of low duty-cycle echolocating bats (Figure 39). Only duration and peak echolocation frequency are reported as indicators of sensory ability of high duty-cycle echolocating bats because the bandwidths of their CF calls are similarly narrow.

Figure 39. A graphical representation of echolocation parameters. Note that these parameters were not measured from the spectrogram: peak frequency and bandwidth were measured from the power spectrum, while duration was measured from the oscillogram (not shown).

      Where possible, genetic analyses confirmed species identification of the bats that were recorded for this book (e.g. Eick et al. 2005, Schoeman 2006, Stoffberg 2007, Stoffberg et al. 2012, Taylor et al.