Qualitative Behaviour Assessment (QBA) emerged in the early 2000s as a way of evaluating the expressive quality of animal behaviour and emotions using qualitative descriptors, such as “playful” or “depressed.” Developed in response to the scepticism of behaviourist attitudes to animal emotions, QBA is now an internationally respected methodology, if still contentious in some circles for what is perceived as an “anthropomorphic” approach. This article results from a research period spent with a UK university laboratory team who were developing species-specific QBA descriptors for the welfare assessment of laboratory mice. The case of the search for a “calm mouse” illuminates the difficulties sometimes encountered in finding the “qualitative” in QBA. It suggests that welfare assessments of animals are epistemologically multiple. Through a historical account of QBA’s emergence, drawing on Cristina Grasseni’s concept of an “ecology of practice,” I argue that different modes of perceiving animal behaviour have emerged through socially and historically inscribed practices.