Animals show strong responses to changes in the risk of being killed by their predators, such as stopping looking for food and being more vigilant. Researchers often study these responses to understand how antipredator behavior has evolved. When predation risk varies how much predators might respond depends on how frequently predators appear and how long they are around for, because they have to make sure that they find enough food when they can forage when the risk is low. In order to interpret experimental results it is important to understand the situations in which responses may be stronger or weaker.
This study used mathematical models to investigate how much animals should respond to predation depending on the frequency and intensity of the risk. The model predicts that animals be more vigilant (and so gather less food) when predation risk is high, and compensate by foraging more at high risk. When high risk is longer or more frequent animals might have no choice but to virtually ignore the risk and forage anyway to avoid starving. When the opposite is true and predators are only around for short infrequent periods animals may be able to stop foraging altogether until the risk has declined.
The common experimental approach of keeping animals under low predation risk and then exposing them to brief signs of predators, such as flying over a silhouette of a predatory bird, may overestimate the responses of animals to predators. In real life animals may not be able to respond so strongly because they need to eat. On the other hand, if vigilance is very costly such as if there are lots of places for predators to hide, previous thinking expected animals could not afford to be too vigilant. However, if these periods are brief and animals manage to mostly forage where they can spot predators easily, they may be able to dedicate lots of time to vigilance during dangerous times. Other reasons for predation risk to vary include the amount of light varying over the day, so animal’s responses to risk are probably very common.