Carotenoid-based colour polyphenism in a moth species: search for fitness correlates

Carotenoid‐based integumental coloration is often associated with individual performance in various animals. This is because the limited amount of the pigment has to be allocated to different vital functions. However, most of the evidence for the carotenoid‐based trade‐off comes from vertebrate studies, and it is unclear if this principle can be applied to insects. This possibility was investigated in Orgyia antiqua L. (Lepidoptera: Lymantriidae). The larvae of this species are polyphenic in their coloration, varying from a highly conspicuous combination of yellow hair tufts on black background to cryptic appearance with brown hair tufts. The conspicuous larvae are aposematic, advertising their aversive hairiness. The maintenance of different colour morphs in O. antiqua requires explanation, as an aposematic signal is expected to evolve towards monomorphism. Chromatographic analysis showed that the yellow coloration of the hair is based on the carotenoid pigment lutein (α‐carotene‐3,3’‐diol). The colour of hair tufts was dependent on their carotenoid content. This justifies an expectation of carotenoid‐based physiological trade‐offs between aposematic coloration and individual performance. To test this hypothesis, we monitored life histories of differently coloured larvae reared on various host plants, recording their body sizes, growth rates, and mortalities in each instar. There was a significant but relatively low heritability of tuft coloration, which allowed us to expect environmental effects. We found no phenotypic associations between hair tuft colour and performance indices in O. antiqua larvae, neither did the quality of host plant affect the frequency of colour morphs. However, the frequency of colour morphs differed between larval instars. Our results suggest that carotenoid‐mediated physiological trade‐offs are not involved in the maintenance of colour morphs in O. antiqua larvae, and factors other than individual condition should be responsible for the observed variability.