Many traits exhibit nonequivalent effects upon maternal versus paternal inheritance. Such “parent-of-origin” effects may be caused by several mechanisms including sex chromosomes and maternal inheritance of mitochondrial DNA. Recently, a class of mammalian autosomal genes has emerged that shows expression of only 1 parental allele. This phenomenon has been termed “genomic imprinting.” Genomic imprinting is an epigenetic effect resulting from chromosomal marks established during gametogenesis. Such imprinted genes result in non-Mendelian inheritance patterns despite being located on autosomes. The chorioallantoic placenta and brain are prominent places of imprinted gene expression. Correspondingly, most imprinted genes appear to be involved in growth or behavior. Interspecific variation in which genes display genomic imprinting suggests that the process is under selection. There is also evidence for intraspecific variation in degree of imprinting of certain genes. Here I briefly review the current understanding of imprinting mechanisms and arguments for selection. The leading argument for positive selection of genomic imprinting is an extension of the concept of kin selection. Although this hypothesis remains controversial, the involvement of imprinted genes in placentation suggests a role in reproductive isolation. Interspecific hybrids in the cricetid genus Peromyscus exhibit parent-of-origin effects involving placental and somatic growth dysplasias. Female P. maniculatus crossed with male P. polionotus produce neonates smaller than either parental strain, with placentas half the parental size. Female P. polionotus crossed with male P. maniculatus produce dysmorphic overgrown embryos whose placentas average >2.5 times the mass of the parental strains. Hybrid dysgenesis in Peromyscus is affected by both the imprinting process and interactions among imprinted genes. I hypothesize that imprinted genes underlie multiple cases of reproductive isolation in the P. maniculatus species complex. Further, I suggest that such interactions have played a significant role in generating mammalian diversity. Finally, I examine the role of the environment in regulating genomic imprinting and argue that studying natural populations in wild-type habitats will be critical to understanding this phenomenon.