The price of safety: food deprivation in early life influences the efficacy of chemical defence in an aposematic moth

Aposematism is the combination of a primary signal with a secondary defence that predators must learn to associate with one another. However, variation in the level of defence, both within and between species, is very common. As secondary defences influence individual fitness, this variation in quality and quantity requires an evolutionary explanation, particularly as it may or may not correlate with variation in primary signals. The costs of defence production are expected to play a considerable role in generating this variation, yet studies of the cost of chemical defence have focused on species that sequester their defences, while studies in species that produce them de novo are scarce. Here we examine effects of resource availability on the production of chemical defences in female wood tiger moths Arctia plantaginis. This species produces defensive fluids when attacked, and advertises this protection with bright colours on its hindwings. Furthermore, the chemicals in these fluids are produced de novo. In order to establish if the production of this chemical protection is costly, and thus resource‐limited, we manipulated resource availability (i.e. food) during larval development and measured its subsequent effects on adult chemical defence. We presented defensive fluids from female moths to wild blue tits, in the absence of any visual cues, to determine their effectiveness against avian predators. Our results demonstrate that the defensive fluids produced by female moths repel birds, and that these defences require resource investment to produce. We found that moths that were resource‐limited during development had less effective chemical defences, despite producing the same volume of defensive fluids as the control moths. As a reduction in available resources negatively influences the effectiveness of the chemical defence, resource availability may explain the variation seen among individuals.