Methodological considerations on the acoustic signal analysis for two species of bats (Chiroptera, Vespertilionidae) [abstract]

C. Zmarich, E. Vernier and F. Ferrero (1997). Methodological considerations on the acoustic signal analysis for two species of bats (Chiroptera, Vespertilionidae) [abstract]. Bioacoustics, Volume 8 (3-4): 275 -276

Ultrasonic emissions of bats (Mammalia, Chiroptera) consist of either social calls or echolocation pulses. As to the latter, every bat species exploits peculiar morphological features in the time and frequency domain, thereby making them distinguishable by their unique echolocation pattern. In this study, the echolocation pulses of two species of vespertilionid bats were recorded twice, first in the laboratory and then in a natural environment. The signals of adult specimens of Pipistrellus kuhlii and Hypsugo savii are analysed and described. These species, which are quite common in Italy, are externally similar and have antropic habits, covering a very much similar ecological role. Their distinction is normally based on areal spreading, physical size and morphology, and acoustic classification criteria. The latter criterion is often used in field recording conditions after heterodyning conversion has made the ultrasonic pulses audible to the human ear. The ultrasonic sounds were checked and detected with a Pettersson Ultrasound Detector D-100 connected to a Schlumemberger magnetic recorder Euromag 1. The recording rate was set to 38 cm/s. As for the laboratory recording conditionss the specimens were recorded while flying inside a 7x4x3 m room. A wide band microphone was set at 1.5 metres from the floor near the central point of the longest wall. As for the field conditions, specimens flying under or near street lamps (P. kuhlii) or over little private gardens in complete darkness/poor light (H. savii) were recorded. In order to apply DSP analysis to the recorded signals, the magnetic tape was played at 4.75 cm/s, making exploitation of the nominal 8 kHz spectrographic standard range possible. In this way duration is extended and frequency is lowered by the same 8-fo1d factor (the virtual interval of 1 kHz thus corresponding to a real interval of 8 kHz). Spectrographic analysis was performed with the DSP Sonograph 5500 and CSL 4300 software by Kay Elemetrics Corp. Measurements for time and frequency were taken through power spectra obtained by positioning the cursor on the waveform at the initial, central and final part of each single signal. Measurements of emission rates were taken simply by counting the occurrences of the homogeneous signal units per second. Some methodological suggestions are made regarding the best way to perform spectrographic analysis on signals that differ in general shape (quasi- constant frequency vs downward frequency modulation) and duration (short vs long). A general difference between the lab and field conditions is the relative greatest length of the signals in the latter. Based on measurement values, it is possible to compare the two species only for the 1ab recordings, because the signals recorded in the field are different for the different species (QCF for H. savii and DFM for P. kuhlii). The DFM signals of the two species both begin at 70 kHz but end at 45 kHz for H. savii and 35 kHz for P. kuhlii. Furthermore, the signals of P. kuhlii have greater amplitude: more steep slope and occur in more rapid succession than those of H. savii.