How does biological diversification occur in the face of genetic exchange? How do reproductive barriers evolve and function? What is the role of adaptive traits in promoting diversification and speciation? These open questions in evolutionary biology are at the core of this project. In order to tackle them, we have focused on butterflies in the neo-tropical genus which are an important component of the diverse butterfly communities in the Neo-tropics. Butterflies in the genus Heliconius are unpalatable to predators, use warning colours to advertise their defences, and mimic other defended butterflies in their local communities. The genus has undergone an adaptive radiation in wing colour patterns as a result of natural selection for mimicry, and is also well known for assortative mating based on wing pattern. I have extended the current knowledge about the ecological function and the genetic basis of wing color patterns in these butterflies and explored the importance of wing coloration relative to chemical signaling in the early stages of diversification. To this aim, I have characterised the adaptive divergence between lineages at different stages of the speciation continuum, by integrating genomic, phenotypic, behavioural, chemical and ecological data. More precisely, I have studied the so-called silvaniform sub-clade of Heliconius, known for harbouring species with tiger patterns that participate in mimicry with large groups of other closely and distantly-related species. My work includes the comparative description of the genetic architecture of wing pattern adaptation in two species, H. hecale and H. ismenius, using crosses, genome-wide next-generation genotyping, and advanced morphometrics of colour patterns. I have also explored the importance of natural and sexual selection on wing-patterning loci at early stages of divergence in the genus. In particular, I have analysed the structure and maintenance of a hybrid zone between two distinctly coloured parapatric races of H. hecale by using a combination of population genetics and genomics, coupled to a phenotypic analysis of the clines and to behavioural assays on male-based mate choice. Finally, I have carried out genome-wide analyses of divergence and gene flow with whole genome sequencing data to look for evidence of introgression between coexisting, hybridising co-mimetic species. This was again coupled to experiments on mating preferences and behavior, and yielded evidence for important differences in putative pheromone signals which may mediate species recognition and the maintenance of species boundaries. Overall, my results show that although selection on wing pattern divergence have been central to the diversification of the genus Heliconius, the accumulation of other barriers to gene flow may be important for the speciation process to be completed.