Porous species boundaries are characterized by differential gene flow, where some regions of the genome experience divergent evolution while others experience the homogenizing effects of gene flow. If species can arise or remain distinct despite gene flow between them, speciation can only be understood on a gene by gene level. To understand the genetics of speciation, we therefore must identify the targets of selection that cause divergent evolution and identify the genetic architecture underlying such “speciation phenotypes”. This will enable characterization of genomic regions that are “free to flow” between species, and those that diverge in the face of gene flow. We discuss this problem in the genus Laupala, a morphologically cryptic, flightless group of crickets that has radiated in Hawaii. Because songs are used in courtship and always distinguish close relatives of Laupala as well as species in sympatry, we argue that songs in Laupala are speciation phenotypes. Here, we present our approaches to identify the underlying genomic regions and song genes that differentiate closely related species. We discuss what is known about the genetic basis of this species difference derived from classic quantitative genetics and quantitative trait locus mapping experiments. We also present a model of the molecular expression of cricket song to assist in our goal to identify the genes involved in song variation. As most species are sympatric and exchange genes with congeners, we discuss the importance of understanding the genetic and genomic architecture of song as a speciation phenotype that must be characterized to identify differential patterns of gene flow at porous species boundaries.