- Bat Call Identification 101
- Licence statement
- Page 2: Horseshoe and Leaf-nosed bats
- Page 3: False Vampires, Free-tailed and Tomb bats
- Page 4: Evening and Bent-winged bats
- Page 5: Concluding remarks, Bibliography
- Bonus: Some bat sounds for your ears!
The extraordinary animals that bats are, they navigate and find their food using a sophisticated SONAR system – a phenomenon called ‘echolocation’. To marvel at an echolocating bat, you need to look no further than your nearest streetlight! Just watch those tiny bats zip, zap and maneuver. Those bats that hunt insects that are attracted to the lights fire ultrasound in all possible directions and wait for the returning echoes. Their waiting time is not in minutes or seconds, but in milliseconds. In those 5-10 milliseconds, these bats make complex mathematical calculations about the position and nature of the prey or the object ahead and make quick decisions that are even faster than a supercomputer. To ecologists, like myself, an additional component of these sounds is interesting. Most of these echolocation calls are species-specific, which means that just like birds, many different species of bats have different calls. But how would you record the ultrasonic calls that you can’t hear? And assuming you do record them, how would you know which species you recorded? The first problem is more tractable. Bats can easily be recorded using ultrasonic recorders called ‘bat detectors’ [watch my video of a bat detector in action here]. There is a growing market (albeit not in India) with a range of options, prices and functionalities. Bat Conservation Trust has a comprehensive article on the different types of bat detectors. The second problem is what this page attempts to solve by presenting a reference library of calls of 57 species of bats found in India (~50% of the remarkable diversity of India’s echolocating bats). Most of the recordings are scientifically curated (peer reviewed) while others are identified with sufficient confidence thanks to the works of fellow bat researchers (see references for details).
Although batwatching, today, is not a popular hobby in India, for the sake of posterity, I have decided to put my recordings out in the public domain, for when the time comes, this page can serve as a standalone reference for your batwatching or research journey. This page is a work in progress. New calls will be added as and when I record them. Alternately, old call IDs will be updated when new (peer-reviewed) information becomes available. I have followed the nomenclature of the latest checklist of bats of South Asia (Srinivasulu et al. 2021, link to the checklist). The nomenclature will be updated following new developments in taxonomy.
Bat Call Identification 101
A typical bird call database such as Xeno-Canto or Macaulay Library has actual recordings of birds that you can listen to. Such a framework doesn’t work for bat calls. Well, one, for obvious reasons because you can’t hear bat calls (with a handful of exceptions). Two, even if you listen to bat calls that are down-sampled to bring them to our hearing range, the differences between two species are barely noticeable. Unlike birds, bats use most of their calls for a completely different purpose ie navigation, orientation and foraging. This predisposes the call to variations with respect to environmental conditions. For example, a species of bat flying in an open space emits long calls because its targets/obstacles are few and far between. The same bat when flying in a forest emits short calls, spanning a wide range of frequencies, to keep track of the countless obstacles in its way.
So, what do you do with a recording that you can’t hear? You see it. Sound can be visualised using a time-frequency (X-Y) graph called ‘spectrogram’. Spectrograms are also loosely called ‘sonograms’ but that is wrong, unless you want to describe the status of your unborn child! A spectrogram allows you to see the structure of a call, its frequency modulation, the duration of each sound pulse and the same between two subsequent pulses. Boring as it may sound, this is what a bat call library is all about: seeing calls and making measurements from a spectrogram.
In order to fully understand my text on bat call ID, it is useful to acquaint yourself with the following terms that describe the call structure and the measurements that help in characterising a bat call. For illustration, here’s a figure from my paper Chakravarty et al. (2020).
- Frequency-modulated (FM, yes, the same one as in a radio!): here, a sound pulse starts at a higher frequency and ends at a lower frequency. Depending on the genus, the pulse the frequency drop can be steep (in a short time interval) or relatively gradual. All bat calls are FM to some degree, but the ‘FMness’ is most pronounced in ‘Mouse-eared’ bats (Myotis spp.) and Tube-nosed bats (Murina spp.).
- Constant-frequency (CF): as opposed to FM, a CF call maintains a constant frequency for much of its duration. FM components may precede and/or succeed the CF part. Examples of CF bats are horseshoe (Rhinolophus spp.) and leaf-nosed bats (family Hipposideridae). Horseshoe bats have an FM-QCF-FM call structure while hipposiderids have CF-FM.
- Quasi-constant-frequency (QCF): in this case, the call shows a shallow frequency modulation, such that it appears largely CF. Free-tailed (family Molossidae) and Tomb bats (family Emballonuridae) use QCF calls. Many evening bats (Vespertilionidae) have calls that start in FM and gradually end as QCF.
- Multi-harmonic: Harmonics, also called ‘overtones’ are a natural element of sound produced by any musical instrument or an animal’s sound production system. I’m not going to explain harmonics here because a) I’m not qualified to teach Physics even to toddlers and b) many people on YouTube have explained it (in more relatable musical contexts) way better than I ever can. Nevertheless, one key property of harmonics is incredibly useful in bat call ID. Harmonics rise in multiples of the fundamental (first) harmonic * (the lowest harmonic that can be recorded). For example, if a bat calls at 10 kHz (first harmonic), its second harmonic will be at 20, third at 30, fourth at 40 and so on. Why is this important? In areas of high bat activity, calls of different species of bats may get stacked one on top of the other. There you need to do your math and check if these are calls from different species or harmonics of the same species. Another reason: harmonics are extremely prominent in the calls of False Vampires (Megadermatidae), Horseshoe (Rhinolophidae), Leaf-nosed (Hipposideridae), Mouse-tailed bats (Rhinopomatidae) and Tomb bats (Emballonuridae). Multiple harmonics are visible in megadermatid calls. However, in the latter three families, usually the second harmonic is dominant (loud and clear on a spectrogram). This property of the emballonurid call is immensely useful in separating them from molossid bats that call at similar frequencies. [*Please note that some people make a distinction between the fundamental and the first harmonic. They consider the fundamental as the lowest, and first as the one above it. Having followed such a debate, I am not sure which is correct but for simplicity I am using ‘first’ and ‘fundamental’ interchangeably.]
- Start frequency (SF): the frequency at which the call starts. In FM bats, the call typically starts at a higher frequency and ends at a lower frequency. This high starting frequency often attenuates when the distance between the bat and the observer (recorder) is large. Therefore, this measurement is highly variable and not very useful for identification.
- End frequency (EF): the frequency at which the call ends. Irrespective of the distance between the bat and the observer, the EF remains well-maintained. In my personal opinion, this is the most useful measurement for identification and I will refer to it throughout the text.
- Peak frequency (PF): The frequency at which the call is the loudest. This is typically measured using a ‘power spectrum’: a graph between frequency and amplitude. But Raven Pro measures it for you directly through the spectrogram. In FM bats, this may be spread across the frequency range and has little value for ID. In FM-QCF bats, this is typically in the QCF component and presents a useful ID trait. In the bat literature this is often referred to as ‘frequency of maximum energy (FmaxE)‘.
- Duration (D): Difference of end time and start time of each pulse. Duration is also measured in other ways (using an ‘oscillogram’) but I believe, this simple method is most easily reproducible.
- Inter-pulse Interval (IPI): The time between the start of one call and the next one. Provided you have a neat call sequence, the IPI is useful in seeing how frequently the bat calls in free-flight (as body size of the bat determines its wing flapping and calling rate). However, IPI varies with habitat (low in clutter, high in open) so make a note of the habitat in your notebooks.
The spectrograms displayed here were analysed using Raven Pro (Cornell Lab of Ornithology, USA), which is freely available to Indian citizens. I made the spectrograms at an FFT size (loosely speaking, the resolution of the spectrogram) of 1024. For best comparisons, I recommend using the same FFT size. The calls are categorised in the captions as follows:
- Release calls: echolocation calls recorded while releasing a bat that was caught and identified at hand. As these bats were handled, their identity is known with greater confidence. The downside, though, is that release calls are not always representative of the calls emitted by a particular species in free-flight (which is how you will encounter them). Release calls give a fair idea of the frequencies at which a given species operates, but do not effectively capture the call structure. Having said so, release calls are still the only way to get a call library started.
- Free-flying: calls recorded of freely-flying bat. These bats were not caught and identified by morphology. The calls were identified based on literature published by myself and other bat researchers from India and abroad. As the species was not seen at hand, these calls must be treated with varying degrees of confidence. Apart from measurements (the objective element of bat call ID), the known geographical distribution, habitat choice and the species’ acoustic behaviour (the subjective element of bat call ID) were taken into account. There might be criticism to my approach but I argue that there is simply no way to arrive at a bat ID (in fact, any animal sound ID) by solely objective means. It is important to use some judgement in identifying bat calls, provided you can explain your reasons – which I do so in the text.
- Feeding buzz: When a freely-flying bat approaches its insect prey, it shortens the duration of each pulse and calls more frequently to keep track of the moving target. Finally, as it jumps into the hunt, its pulses are so close together that they sound like a buzz. This is called a ‘feeding buzz’ and it allows one to know whether the bat is actually hunting in a given area or merely moving around.
- Social calls: Bats use a variety of calls for communicating with their peers. These calls come in the form of chirps, yodels, whistles, shrieks, buzzes or hums (among others). Some calls are audible to the non-hearing-impaired and others can be heard for their melodious properties by down-sampling. Unfortunately, all of this action happens in pitch darkness so we don’t know the context behind most calls. This compels bat researchers to lump all these communication calls under the broad umbrella of ‘social calls’. Social calls often fall in between echolocation calls when the bat is in flight. This way, the social call be assigned to the given species. However, in areas of high bat activity, this may be prone to error. Social calls can be advertised from an ultrasound speaker to attract bats.
- Distress calls: These are calls that bats emit when being handled. Like social calls, they can be used to attract bats using call playback.
Licence statement/Terms and conditions of usage
- Sharing and non-commercial projects: All the spectrograms on this page can be shared and be used in non-commercial projects (research papers, reports, magazine articles etc) with due credits (preferably with a link to the webpage). Should you need a higher resolution, please write me an email.
- Research projects: The original .wav files with associated metadata are available upon request provided you explain the purpose. If you need the files for a research project and believe that these files and/or my contribution would hold value in your research, I would be happy to collaborate. Please write to me at email@example.com
- Books: The spectrograms cannot be reproduced in books without my permission. Please write to me at firstname.lastname@example.org.