Adaptive radiation is recognized by a rapid burst of phenotypic, ecological, and species diversification. However, we still have no quantitative model of this common process and a poor understanding of the topography of fitness landscapes underlying this process for nearly all natural populations. I developed a case study of adaptive radiation in Caribbean pupfishes, integrating phylogenetic comparative methods, population genomics, field measurements of natural selection, functional morphology, and behavioral ecology. Here I discuss how empirical measurements of the fitness landscape in the wild, measured from the growth and survival of 2,000 F2 hybrids placed in field enclosures, help explain these patterns of rapid diversification, adaptive radiation, and the origins of novelty.