Speciation is a central mechanism of biological diversification. While speciation is well studied in plants and animals, in comparison, relatively little is known about speciation in fungi. One fungal model is the Cryptococcus genus, which is best known for the pathogenic Cryptococcus neoformans/Cryptococcus gattii species complex that causes over 200,000 new infections in humans annually. The closest non-human pathogenic relatives are the sibling species, Cryptococcus amylolentus and Tsuchiyaea wingfieldii. However, because relatively few isolates of each species are available, it is unclear whether they represent divergent lineages of the same species or different biological species. The recent isolation of an additional strain, preliminarily identified as T. wingfieldii, prompted us to reexamine this group as it may inform about the evolutionary processes underlying the diversification of both non-pathogenic and pathogenic Cryptococcus lineages. Using genomic data, we reappraised the phylogenetic relationship of the four available strains and confirmed the genetic separation of C. amylolentus and T. wingfieldii (now Cryptococcus wingfieldii), and revealed an additional cryptic species, for which the name Cryptococcus floricola is proposed. Comparison of full-length chromosome assemblies revealed approximately 6% pairwise sequence divergence between the three species, and identified significant genomic changes, including inversions as well as a reciprocal translocation that involved inter-centromeric ectopic recombination, which together likely impose significant barriers to genetic exchange. Using genetic crosses, we show that while C. wingfieldii cannot interbreed with any of the other strains, C. floricola can undergo sexual reproduction with C. amylolentus. However, most of the spores resulting from this cross were inviable, and many were sterile, indicating that the two species are genetically isolated through intrinsic post-zygotic barriers and possibly due to niche differentiation. Genome sequencing and analysis of the progeny demonstrated decreased recombination frequency during meiosis in heterospecific crosses compared to C. amylolentus conspecific crosses. This study advances our understanding of speciation in fungi and highlights the power of genomics in assisting our ability to correctly identify and discriminate fungal species.Author SummaryThe idea of species as discrete natural units seems rather intuitive for most people, just as cells are the basic units of life. However, when observing variation across a species range, boundaries can become blurred making it less than obvious when different populations evolve into separate species. Additionally, separate species can still interbreed, such as lions breeding with tigers to produce a liger or a tigon (depending on the paternal and maternal species of origin), but the resulting offspring is usually inviable or sterile, which in turn is evidence that the parents involved are distinct species. Therefore, what species are and how they originate is still an open question in evolutionary biology. While recent advances have been made in the fields of animal and plant speciation, many other important components of biological diversity, such as fungi, are still understudied. Genome sequencing is now providing new tools to address the genetic mechanisms that drive divergence and reproductive isolation between populations, including genetic incompatibilities, sequence divergence, and chromosomal rearrangements. Here we focus on the Cryptococcus amylolentus species complex, a non-pathogenic fungal lineage closely related to the human pathogenic Cryptococcus neoformans/Cryptococcus gattii complex. Using genetic and genomic analysis we reexamined the species boundaries of four available isolates within the C. amylolentus complex and revealed three genetically isolated species. The genomes of these species are ~6% divergent and exhibit chromosome rearrangements, including translocations and small-scale inversions. Although two of the species (C. amylolentus and newly described C. floricola) are still able to interbreed, the resulting hybrid progeny were mostly inviable, and many were sterile, indicating that barriers to reproduction have already been established. Our results will foster additional studies addressing the transitions between non-pathogenic and pathogenic Cryptococcus lineages.