This review evaluates direct (live-trapping) and indirect (genetic) methods to study dispersal in wild house mice (Mus musculus) and summarizes field and experimental data to examine the causes and consequences of dispersal. Commensal house mice (associated with human habitations, farms, food stores and other anthropogenic habitats) typically show lower rates of dispersal than feral house mice (living in crops, natural and semi-natural habitats). However, early claims of long-term fine-scale genetic structure in commensal house mice (due to low rates of dispersal) are not supported by recent data. Dispersal becomes obligatory when habitat conditions deteriorate, but most dispersal occurs below the local environmental carrying capacity and is due to social interactions with conspecifics. Excursions are relatively frequent and probably allow mice to assess the quality of habitats before dispersing. Young males have the greatest tendency to disperse, apparently prompted mainly by aggressive interactions with dominant males. If they do disperse, females integrate into new groups more easily than do males. Dispersing house mice risk loss of condition or death, but may gain reproductive opportunities on arrival in a new location. House mice can be transported passively as stowaways with humans; this contributes to population persistence and genetic structure at regional scales and has allowed house mice to spread world-wide. © 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 84, 565–583.