The disjunct distribution of the quokka enabled this study to investigate cranial morphological variation in relation to insularity and latitude. Crania from mainland locations in south-western Australia and from two islands were examined. Thirty-eight three-dimensional homologous landmarks were digitized on 110 quokka crania. The landmark data were first subjected to generalized Procrustes Analysis, followed by principal components analysis. General linear regression was used to test whether quokka cranial size obeys the Island rule and Bergmann's rule. Crania from the islands were found to be smaller than crania from the adjacent mainland. Crania of quokkas from cooler climates (higher latitudes) were larger than those from warmer climates (lower latitudes), and this trend was observed in both mainland and island samples. Multivariate regression of the combined principal components on the independent variable, latitude, showed shape of the snout of quokkas from cooler southern climates tended to be relatively longer and narrower than those from further north. General linear regressions was also used to ensure that latitude was significantly influencing shape (principal components) independent of age, sex and whether the crania were from the mainland or island populations. Results showed that 40.2% of cranial shape variation was significantly related to latitude, irrespective of age, sex and population type. The variation in size and shape of the quokka crania appears to reflect the eco-geographic variation within the distribution of the species.