We went to the Peruvian Amazon in order to complete this project on the study of the butterfly Heliconius numata. This toxic butterfly always displays a high diversity of predator-warning colour patterns within populations, with up to seven distinct morphs coexisting within a given locality. Each one of these morphs is a perfect mimic of different species of butterflies flying in the same localities. This perplexing warning colour diversity is under the control of a single supergene locus, and alleles associated with distinct morphs are characterised by different combinations of chromosomal inversions.
Using more than 8000 artificial butterflies displaying colour pattern elements exhibited by Heliconius numata, we tested the benefit of inversions in capturing co-adapted combinations of alleles directly within the natural habitat. We revealed that combinations of wing elements naturally found in H. numata and which result in a mimetic appearance offers very strong protection from predators. In comparison, combinations of wing colour pattern elements resulting in a non-mimetic appearance, as expected if genetic recombination was not suppressed by inversions, suffered high rates of predation. Such drastic differences in the levels of protection provided to these two types of equally bright warning phenotypes highlights the adaptive potential of chromosome inversions in locking together colour elements resulting in a beneficial mimetic phenotype. We show that this selection for mimicry stems from the avoidance learning behaviour of avian predators toward warning signals of unpalatability. Indeed, such behaviour generates natural selective pressures that are characterised by the efficiency of a warning signal increasing with its local abundance within natural habitats. Because recombinant phenotypes do not benefit from the advantage of a high abundance in combination with other mimetic species of the community, they are highly predated.
Nonetheless, the diversity of warning signals observed in H. numata is unexpected because selection by predators favours the most abundant morph. The coexistence of distinct phenotypes within a locality involves the maintenance of one well-protected form and a diversity of rarer forms which suffer more predation events. By investigating sexual preference in this species, through extensive mate choice experiments, we have shown that female choice drives a mating strategy favouring the pairing of dissimilar forms and promoting allelic diversity at the mimicry supergene. Overall, we have shown that the warning signals of H. numata are subject to two antagonistic selection regimes: sexual selection promoting polymorphism, and purifying selection by predators promoting the fixation of the most abundant mimetic wing pattern.
By maintaining suboptimal morphs, the evolutionary significance of such sexual behaviour remains difficult to comprehend. Survival studies of the distinct inversion genotype at the mimicry supergene assessed through experimental H. numata crosses however, have revealed the evolutionary benefit of maintaining multiple mimetic morphs in sympatry. Indeed, we show that the physical properties of the chromosome inversion, which is at the heart of the beneficial mimetic phenotypes, causes a lethal recessive genetic disease that is expressed in homozygote individuals. By minimising the frequency of homozygotes for the inversion in offspring, disassortative mating may have evolved in response to the deleterious effects associated with the inversions.
In conclusion, this project provides a comprehensive case example of how adaptive polymorphism may result from a trade-off between the beneficial effects of co-adapted mutations (involved here in mimicry adaptation and the formation of a supergene), and deleterious effects associated with the physical properties of inversions (associated here with certain supergene variants).
