Over the years, the evolutionary importance of natural hybridization has been a contentious issue. At one extreme is the relatively common view of hybridization as an evolutionarily unimportant process. A less common perspective, but one that has gained support over the past decade, is that of hybridization as a relatively widespread and potentially creative evolutionary process. Indeed, studies documenting the production of hybrid genotypes exhibiting a wide range of fitnesses have become increasingly common. In this review, we examine the genetic basis of such variation in hybrid fitness. In particular, we assess the genetic architecture of hybrid inferiority (both sterility and inviability). We then extend our discussion to the genetic basis of increased fitness in certain hybrid genotypes. The available evidence argues that hybrid inferiority is the result of widespread negative epistasis in a hybrid genetic background. In contrast, increased hybrid fitness can be most readily explained through the segregation of additive genetic factors, with epistasis playing a more limited role.