<p>The adaptive function of banded and other repetitive contrasted patterns in insects is not fully explained; they are often proposed to be aposematic, but the multiple independent evolution of certain complex patterns and the similarity of patterns between non-defended and toxic caterpillars suggest the possibility of another selective force. Zebras’ stripes are no longer proposed to be aposematic, but have been shown to have the function of deflecting biting flies; stripes interfere with flies’ ability to navigate near zebras and improve fitness potentially by increasing the efficiency of zebras’ primary defence against flies: biting, tail-switching and escaping. I propose that, as in zebras, repeated contrasted patterns evolved in caterpillars to interfere with flies’ visual systems and increase their fitness by improving the efficiency of their behavioural primary defences. This inference is supported by the empirical precedent of optical-interference-based protection against flying insects in zebras, and by the intense selection exerted on caterpillars by flying insects, notably parasitoids, major natural enemies of many caterpillars, particularly toxic ones. The intense selection by parasitoids appears to have selected for toxicity and behavioural primary defences such as appendage switching and biting. The distances at which caterpillar bands are predicted to impair flies are short, and such impairment could increase the efficiency of these defences. This selective force would explain the convergence in patterns across species and continents, and, unlike aposematism, can be reconciled with the evolution of (non-mimicry) high contrasts in edible caterpillars and the association of polymorphism with contrasts in caterpillars.</p>