Fusarium oxysporum is a cross-kingdom pathogenic fungus that can cause vascular wilt disease in many economically important plants and local or disseminated infections in humans. Although it lacks a sexual stage in its life cycle, F. oxysporum can adapt to a wide range of hosts because of accessory chromosomes (ACs) which are enriched in host-specific genes and repeat content. This dissertation investigates the mechanisms that drive the adaptive evolution in the cross-kingdom pathogen F. oxysporum using comparative genomics and an experimental evolution approach. The first chapter compares phenotypes and genomes of a plant pathogenic isolate F. oxysporum f. sp. lycopersici 4287 (Fol4287) and a human pathogenic isolate F. oxysporum MRL8996. Fol4287 and MRL8996 differ in both morphology and AC gene and repeat content. The second chapter analyzes the Fol4287 populations generated by evolution experiments through whole-genome sequencing. Many copy number variations, single nucleotide variations, insertions deletions, and transposable element insertion variations (TIV) were detected. The final chapter compares the adaptive mutation mechanisms in Fol4287 and MRL8996 through experimental evolution. While transposons are the major cause of variation in both strains, the active transposons are different and encoded in their ACs. For Fol4287 a DNA transposon, Hormin is highly active while in MRL8996, a short interspersed nuclear element, Foxy5, had the highest activity. The dynamic chromosomes are also different in both strains: in Fol4287 ACs and chromosome 13 have many copy number variations while in MRL8996 ACs are relatively stable while chromosome 12 is unstable. In addition, the Velvet complex, which is a major regulator of growth and mutated in multiple independent populations has an important role in the adaptation of F. oxysporum. Many interesting evolutionary events are also observed. In conclusion, the accessory chromosomes provide evolutionary hotspots for F. oxysporum.