Humans and other animals with a big X and a small gene-poor Y chromosome share the problem that many X-borne genes are present in two copies in females and a single copy in males. In mammals, compensation for this different gene dosage is accomplished by inactivation of one X in the somatic cells of females. Discovered by Mary Lyon in 1961, X chromosome inactivation involves the silencing of a thousand unrelated but physically linked genes on one X. It is a whole-X event, involving major cytological changes, including late DNA replication and visibly different compaction into ‘sex chromatin’. Some genes on the X (many in humans, few in mice) escape inactivation. Inactivation is a stable change, inherited by somatic cells but reversed in germ cells, and seems to be controlled from an inactivation centre that can be mapped on the X. X inactivation occurs in the embryo, silencing one or other X at random, but imprinted X inactivation of the paternally derived X occurs at early developmental stages in mice, and at all stages in marsupials. X inactivation is a spectacular example of ‘epigenetic’ silencing on a grand scale, and is intensively studied in humans and mice, and modelled in stem cells.