The dogma that adult tissue-specific stem cells remain committed to supporting only their own tissue has been challenged; a new hypothesis, that adult stem cells demonstrate plasticity in their repertoires, is being tested. This is important because it seems possible that haematopoietic stem cells, for example, could be exploited to generate and perhaps deliver cell-based therapies deep within existing nonhaematopoietic organs. Much of the evidence for plasticity derives from histological studies of tissues from patients or animals that have received grafts of cells or whole organs, from a donor bearing (or lacking) a definitive marker. Detection in the recipient of appropriately differentiated cells bearing the donor marker is indicative of a switch in phenotype of a stem cell or a member of a transit amplifying population or of a differentiated cell. In this review, we discuss evidence for these changes occurring but do not consider the molecular basis of cell commitment. In general, the extent of engraftment is low but may be increased if tissues are damaged. In model systems of liver regeneration, the repeated application of a selection pressure increases levels of engraftment considerably; how this occurs is unclear. Cell fusion plays a part in regeneration and remodelling of the liver, skeletal muscle and even regions of the brain. Genetic disease may be amenable to some forms of cell therapy, yet immune rejection will present challenges. Graft-vs.-host disease will continue to present problems, although this may be avoided if the cells were derived from the recipient or they were tolerized. Despite great expectations for cellular therapies, there are indications that attempts to replace missing proteins could be confounded simply by the development of specific immunity that rejects the new phenotype.