Biological diversity is created by the formation of new species and thus speciation is a fundamental process responsible for diversity of life. Progress has been made in identifying individual ‘speciation genes’ causing specific forms of reproductive isolation. In contrast, debates persist about how the speciation process unfolds from beginning to end (i.e., the mechanisms of speciation) and the fraction of the genome involved. What factors explain variability in how far speciation proceeds? This question was framed by Darwin and an active debate concerns whether speciation proceeds further via: (1) strong natural selection on a few genes, resulting in isolated ‘genomic islands of divergence’, or (2) by selection acting on a greater number of genes across the genome. This proposal outlines a research program aimed at resolving this debate, a task that is required to understand the origins of diversity. Specifically, I will use field experiments which measure natural selection at the genomic level to directly test, for the first time, how selection acts on the genome across different stages of the speciation process. The approach uses an integrative combination of eco-morphology, next-generation DNA sequencing, population genomics, manipulative experiments in nature, and linkage mapping to study the origins of diversity. The work is novel in: (1) examining many closely-related taxon pairs (rather than just a few) which vary in how far speciation has proceeded, thereby reconstructing how speciation unfolds, and (2) experimentally measuring (rather than inferring via observation) selection on the genome. The proposed work will clarify our understanding of speciation. By coupling genomics and ecology via experimental approaches, the results will build a new understanding of the natural history of the genome.