Aim. To identify Antarctic palaeoendemic taxa and their probable glacial refugia from regional groups of endemic species records. Location. Antarctica. Methods. We compiled a list of Antarctic non-marine invertebrates from published literature, and then deleted all records relating to non-endemic, zoochoric (phoretic and parasitic), marine and partially identified species to leave only the elements endemic to Antarctica. We then used cluster analysis and principal components analysis to identify regional groupings within this endemic fauna. Results. Some 170+ of the reported 520+ Antarctic invertebrates are free-living and endemic, but only nine of these are pan-Antarctic, with the majority having either ‘continental’/eastern or ‘maritime’/western distributions. Main conclusions. All invertebrates endemic to continental Antarctica are confined to, or found adjacent to, ice-free palaeorefugial mountains, nunataks and coastal exposures. By contrast, only one maritime Antarctic palaeorefugium has been identified, and most endemic taxa are currently associated with coastal lowland neorefugia. We suggest which regions of Antarctica (1) are likely to be refugial, and (2) simply require more data in order that the nature and origin of their fauna can be elucidated.