An investigation into multi-component warning colouration

Butterflies and moths present conspicuous colour patterns that signal toxicity to predators. The existence of redundancies (warning signals on different body parts) in conspicuous signals is a theoretical conundrum, as they are thought to interfere with predator avoidance learning. Here, we investigate warning colour components in Australian Lepidopterans using three approaches. First, we scored the absence and presence of redundant warning signals across Australian lepidopterans. Then, we reviewed warning colouration methodologies to assess the appropriateness of frequently used quantification techniques. We implemented this knowledge in the final phase, through quantifying warning signal size in both static and dynamic contexts. Redundant signals (warning colours on both wing and abdomen) were found in ~44% of warning-coloured moths, and ~32% of warning-coloured butterflies. Literature analysis revealed that warning signals are generally quantified in a static context, despite broad acknowledgment that signalling is dynamic. Our novel signal-measuring method, tested on the Amata nigriceps moth, confirmed that context (static and dynamic) affects signal appearance, and that one signal element can receive higher selection pressure than another in a colouration pattern. Our results may explain the persistence of overall signal variation, due to the relaxed selection of specific components, as well as redundant signals.