Opinions are divided on the question whether the notion of "natural kinds' plays an important role in philosophy of science. Some authors have argued that it does not, because the idea of natural kinds features only in the early stages of the development of scientific fields of work and ceases to play a role once the field becomes established. Others have argued that it does not, because natural kinds are found only on the most fundamental levels of organization, so that most scientific fields of work do not study natural kinds. Against this dim view of the importance of the notion of "natural kinds' for philosophy of science, I shall argue that it roots in a particular tradition of thinking about the problem of natural kinds and that there is an alternative way of thinking about natural kinds that does not give rise to such a dim view. In so doing, I shall explore what the principal challenge is for philosophers who attempt to devise an account of natural kinds.