The role of natural selection in driving speciation between interbreeding populations has been one of the most controversial topics in evolutionary biology. Populations that inhabit contrasting environments can evolve adaptive traits in the local habitat that can reproductively isolate them. This process seems straight forward between allopatric populations, where populations isolated by geographic barriers can accumulate adaptive genetic differentiation and evolve reproductive isolation, but when populations are in the same or proximate localities, gene flow may oppose divergence and speciation. Although speciation with gene flow is an increasingly supported phenomenon, studies still face complications dissecting the effect of divergence time vs. gene flow, even with extensive molecular data. An increasing number of examples from nature provide evidence for this model of speciation. However, most of the studies have been conducted in animal systems, leaving ecological speciation in plants largely unexplored. Dune and Headland populations within the Senecio lautus ecotype and species complex occur proximate to each other in several coastal localities of Australia, displaying very contrasting morphologies despite being interfertile. Supported by a robust phylogenetic study, each Dune and Headland pair shows an independent origin and it displays characteristics that suggest an important role for ecology in the diversification of its ecotypes. Here, I used a combination of ecological, molecular and comparative approaches to investigate the process of speciation in the Australian groundsel Senecio lautus. In reciprocal transplants in the field and experiments in the glasshouse I found that Dune and Headland populations are strongly isolated by ecology based reproductive barriers and that intrinsic barriers contribute little to it. Then I discovered that most parapatric pairs displayed drastic reductions in gene flow, while more distant populations from the same ecotype still exchange genes. Finally, I provide evidence that the multiple Dune and Headland pairs in the system are evolving under the model of ecological parallel speciation thus providing strong evidence for the role of natural selection in plant speciation. The experimental results in my dissertation suggest that ecology is not only able to counteract gene flow at early stages of divergence, but that can also take populations to the most advanced stages of speciation where populations no longer exchange genes in the field. These results strengthen previous studies that suggest that the evolution of intrinsic reproductive isolation may be decoupled from the process of speciation. Senecio lautus constitutes the first well-supported case for the parallel ecological speciation in plants, and provides an excellent opportunity to study speciation with gene flow.