Antipredator responses to alarm pheromone in groups of young and/or old thrips larvae

Abstract

Many prey species suffer from different predators in the course of their ontogeny. Hence, the alarm signal a small prey individual sends can have a different meaning than the signal a large prey individual sends, both for small and for large receivers. Larvae of Western Flower Thrips face predators that attack only small larvae, or predators that attack small larvae and large larvae. Furthermore, thrips larvae release a two‐component alarm pheromone, which varies in composition with larval age. Here, we study whether their response to alarm pheromone varies with composition of the pheromone. First, we confirmed that large and small larvae respond when nearby larvae of both sizes were prodded with a brush to induce alarm pheromone excretion. Subsequently, we tested whether thrips larvae of a given size respond differentially to alarm pheromone excreted by a small or large companion larva. We analyzed two types of behavior used in direct defense against a predator and one type of escape response. Only small (not large) larvae attempted to escape more frequently in response to excretions from a large larva. This difference in response could have been due to the alarm pheromone or to the companion larva in the vicinity. We subsequently tested for, but did not find, an effect of size of the companion larva on the behavior of the test larva when exposed to synthetic pheromone mimicking that of a large larva. Finally, we tested how pheromone composition affects antipredator behavior by exposing thrips larvae to synthetic pheromones differing in amount and ratio of the two components. Only for small larvae, we found significant changes in escape behavior with pheromone amount, and a trend with the ratio. Overall, we conclude that small thrips larvae respond differentially to alarm pheromones excreted by small and large larvae and that this differential response is due to differences in pheromone quantity and possibly also quality. Our results suggest that responses to alarm signals can vary with the chemical composition of those alarm signals.