This paper focuses on Thomas S. Kuhn's work on taxonomic concepts and how it relates to empirical work from the cognitive sciences on categorization and conceptual development. I shall first review the basic features of Kuhn's family resemblance account and compare to work from the cognitive sciences. I shall then show how Kuhn's account can be extended to cover the development of new taxonomies in science, and I shall illustrate by a detailed case study that Kuhn himself mentioned only briefly in his own work, namely the discovery of X-rays and radioactivity.