Genomic imprinting is an important and enigmatic form of gene regulation in mammals in which one copy of a gene is silenced in a manner determined by its parental history. Imprinted genes range from those with constitutive monoallelic silencing to those, typically more remote from imprinting control regions, that display developmentally regulated, tissue-specific or partial monoallelic expression. This diversity may make these genes, and the processes they control, more or less sensitive to factors that modify or disrupt epigenetic marks. Imprinted genes have important functions in development and physiology, including major endocrine/neuroendocrine axes. Owing to is central role in coordinating growth, metabolism and reproduction, as well as evidence from genetic and knockout studies, the hypothalamus may be a focus for imprinted gene action. Are there unifying principles that explain why a gene should be imprinted? Conflict between parental genomes over limiting maternal resources, but also co-adaptation between mothers and offspring, have been invoked to explain the evolution of imprinting. Recent reports suggest there may be many more genes imprinted in the hypothalamus than hitherto expected, and it will be important for these new candidates to be validated and to determine whether they conform to current notions of how imprinting is regulated. In fully evaluating the role of imprinted genes in the hypothalamus, much work needs to be done to identify the specific neuronal populations in which particular genes are expressed, establish whether there are pathways in common and whether imprinted genes are involved in long-term programming of hypothalamic functions.