Grasshoppers do not flee from their alarmed predators but from non-alarmed ones

Alarm calls produced by basal prey have well-known informative value. In multipredator communities, mesopredators, when faced with top predators, may emit alarm calls that could inform basal prey about their lowered predation risk. To test this unexplored possibility, we conducted one field and one mesocosm experiment in which we simulated alarm and non-alarm calls from little owls (Athene noctua) as mesopredators and measured their effects on grasshoppers as prey of little owls but not of top predators. In the field experiment, we found that grasshopper species were significantly more abundant in patches where we simulated either the presence of scared little owls (alarm treatment) or no owls (control treatment) compared to patches where the presence of non-scared little owls (non-alarm treatment) was simulated. In the mesocosm experiment, locusts’ (Locusta migratoria) moved significantly more to exposed areas when we simulated the presence of scared little owls (alarm treatment) or of a granivorous bird (control treatment), while they moved to sheltered areas when we simulated the presence of non-scared owls (non-alarm treatment). These results show that prey could cue on predators’ calls to assess their predation risk and make decisions, revealing unprecedented potential ecological consequences of alarm calls in invertebrate communities.

Methods
Firstly, in a field experiment we sampled the abundance of several grasshopper species in patches after broadcasting repeatedly for several days either little owls’ alarm calls (alarm treatment, simulating scared owls) or non-alarm/contact calls (non-alarm treatment, simulating non-scared owls), or just visiting the patches (control treatment, simulating no owls). In addition, in a mesocosm experiment, we recorded the distribution of migratory locusts (Locusta migratoria) in exposed and sheltered areas before and after broadcasting little owls´ alarm calls (alarm treatment), non-alarm/contact calls (non-alarm treatment) or vocalizations of a woodpigeon (Columba palumbus) as a granivorous, non-predator bird for grasshoppers (control treatment).

The abundance of grasshoppers in the field experiment were sampled in all patches by counting all detected individuals, either inactive or during their flights/jumps, while walking at constant speed (about 2 km per hour) along longitudinal transects 50 m long and with a 5-m band. The number of grasshoppers was used as the response variable to the call treatment because it is a proxy of insect abundance as prey of little owls.

In the mesocosm experiment, locusts behavior was recorded with video cameras and we counted the proportion of individuals in exposed areas at the end of pre-treatment and treatment conditions (day 1st and 2nd of the trials) and calculated the difference as a measure of preference to be exposed versus hidden in response to the call treatments.