Methodology Description
Playback experiments: The first source of data was from recordings of unhabituated Olive colobus monkeys residing with Diana monkey groups that were exposed to playbacks of predator vocalizations (leopard growls, eagle shrieks). It was compiled between 1994 and 1999 administered to N = 11 vocally responsive groups, some of which were tested with more than one stimulus. The second source of data was also from unhabituated groups administered to N = 18 responsive groups between 2021 and 2022. These groups were exposed to leopard growls, eagle shrieks, chimpanzee pant-hoots, and the sounds of falling trees. For both sources, we used recordings of leopard growls purchased from the British Library of Wildlife Sounds (BBC master tape number MM35; South African Broadcasting Corporation), whereas eagle and chimpanzee vocalizations and falling tree sounds were recorded directly in the study area. Some additional falling tree sounds were purchased from the BBC Sound Effects library (sounds 07002293, 07058030, 07058129). To avoid pseudo-replication, different versions of each stimulus type were produced (N = 4 eagle shrieks of 15 s, N = 4 leopard growls of 15 s, N = 2 chimpanzee pant-hoots of 15 s, and N = 7 falling tree sounds of 9 or 15 s) so that a particular group only contributed with one reaction per version. Sounds were broadcast with a NAGRA DSM speaker, typically positioned near the ground or in the canopy (up to 30 m: eagle and leopard playbacks only). We adjusted the speaker volume so that all stimuli were broadcast with a naturally sounding range. Predator model experiments: We also recorded alarm calls given in response to visual predator models. Here, we tested N = 8 unhabituated groups. Again, groups were located, monitored and recorded as before. We used both a leopard model, which consisted of an experimenter covering himself with leopard-patterned fabric mimicking the size, shape and posture of a leopard and a crowned eagle model, which consisted of an experimenter hiding behind a camouflage net displaying a life-size 3D printed and hand-painted crowned eagle (see SWISSUbase: https://doi.org/10.60544/g1sx-4j76 for 3D print file). In both cases, the experimenter slowly approached a subject in the periphery of the group. All models were presented in motion to facilitate detection. After detection, the predator model remained in sight of the focal subject for a maximum of 1 min before slowly moving away and hiding behind a tree trunk. To address issues with pseudo-replication, we presented two different versions of each predator model to monkey groups. For all data collection period, we systematically searched the study area until a mixed species group was found, typically by hearing Diana monkey or red colobus monkey vocalizations. We always tried to remain undetected by avoiding visual contact with the group, and monitored their behavior from a distance of about 25–75 m for at least 15 min prior to initiating a playback trial. Since we worked with unhabituated groups whose home ranges were unknown, to avoid retesting a group with the same stimulus more than once, we did not test any group in an area of 1 km around the location of the experiment for at least one week (Olive colobus mean home range = 0.56 km2, corresponding to an 850-metre diameter disc), and we ensured that each stimulus was never presented more than once at the same location.
Equipment Used
We played back sounds from a Samsung Galaxy XCover 4 (model SM-G390F, with media sound on maximum volume) connected to a Nagra Kudelski DSM-monitor loudspeaker (frequency response 60– 15,000Hz ± 4 dB), from an iPhone 5S (iOS 12.5.4) connected to an Alpha speaker (AER, The Acoustic People: frequency range 60–18,000 Hz), or from a Sony WMD6C Professional Recorder connected to a Nagra Kudelski DSM-monitor loudspeaker. We recorded vocal responses with Sony WMD6C or TCM5000EV recorders and Sennheiser ME88 or ME67 directional microphones (frequency response, 40–20,000Hz ± 2.5 dB). We also recorded responses with a Marantz PMD 661 MKII recorder and a Sennheiser MKH 416 P48 directional microphone (frequency response, 40–20,000Hz ± 2.5 dB). 1994–99 recordings were digitized with Audacity software v2.1.0 (44.1 kHz sampling rate, 16 bits accuracy, WAV format) using TASCAM CD-A500 and Technics M280 devices.