X chromosome inactivation represents one of the most dramatic examples of mono-allelic gene expression and long-term gene-silencing in mammals. The key regulatory molecule that triggers silencing is the Xist transcript, but little is known about its repressive action. Some progress has been made in deciphering the epigenetics of the inactive state that it triggers, however. During pre-implantation development, the inactive state is relatively labile. Later on, in the soma, the inactive state is highly stable and clonally heritable. This is ensured by the panoply of epigenetic modifications that characterize the inactive X and, presumably, is also a result of its spatio-temporal segregation. The inactive X chromosome has been associated with an increasing number of histone modifications, and several recent studies have implicated Polycomb group proteins in laying down some of these marks. Thanks to genetic and biochemical approaches to analyse these proteins, the epigenetic tapestry of the inactive X is just beginning to be unravelled. Lineage-specific differences provide a glimpse into the developmental complexity of the epigenetic marks that ensure the inactive state.