We tested the hypothesis that undefended home ranges are larger than defended home ranges using data collected from the literature for three groups of mammals. A matched-pairs analysis of populations within species or species within genera showed that undefended home ranges were larger than defended home ranges for carnivores and male ungulates, but not for primates. Primates may have been an exception because they violated a key assumption of the hypothesis, that defence costs increase with the size of the defended area. Undefended home ranges were 5.4 and 15.2 times larger than defended home ranges for carnivores and male ungulates, respectively. Whether or not a home range is defended is an important source of variation that should be included in future studies of home range size.