Abstract Although mammalian tail length relative to body length is considered indicative of locomotor mode, this association has been difficult to quantify. This could be because the counterweight function of the tail might associate it more with body weight than body length. Alternatively, relative tail length might not be evolutionarily flexible owing to its integration with the remaining skeleton, particularly the spine. Using comparative analyses of morphological means and ranges in Australian marsupials, including the first co-assessment with body weight, our study supports the second hypothesis, i.e. tail length ranges within species, and tail lengths among species are explained better by body length than by body weight. However, all three variables do not differ in phylogenetic signal or rates of evolution. Associations of tail lengths with locomotion are limited, but suggest that scaling slopes, rather than intercepts, are responsible for limited divergence between relative tail lengths at different locomotor modes. This complicates (palaeo-)ecological interpretations of tail length further. We conclude that relative tail length is not a strong predictor of locomotor mode, probably owing to strong integration of tail and body length. The many well-documented bony and soft-tissue adaptations of tails are likely to be better suited to interpretations of locomotor adaptations.