Interspecific communication in African savannah herbivores

The study was conducted in the Masai Mara National Reserve as part a joint study between the University of Liverpool and the University of York. Field data were collected by experienced observers following standardised protocols. Vigilance data were video-recorded and vigilance rates subsequently analysed in BORIS, the Behavioural Observation Research Interactive Software. Responses to predator dummies were categorized based on the occurrence of alarm calls within 5 minutes of predator detection. Playback experiments were video-recorded and analysed in BORIS using a clearly defined behavioural protocol. In each case, we excluded experiments that have been disturbed by approaching predators, humans or alarm calling heterospecifics (including alarm calling conspecifics in vigilance recordings). Species counts were conducted along a pre-defined track on three different study plains. Again, we excluded sightings of groups with predators close by. The full description of the methodology is available in the metadata file submitted with this data set. Data were entered within 48 hours of recording and checked for accidental typos and unusual values.