Abstract.1. Two usual definitions of host discrimination are presented. The first is ‘the ability of a parasite to distinguish unparasitized from parasitized hosts and to lay eggs in the former’. This definition is not useful and even confusing since it does not include the ability of a parasite to distinguish hosts containing different numbers of parasite eggs and to lay preferably only in those with the lowest numbers.2. The second definition is ‘the ability of a parasite to distribute its eggs in a non‐random, regular way among its hosts’. It is argued that most field data are insufficient to permit any conclusions about host discriminative ability on the basis of this definition. An example is given of an apparent random distribution by parasites that are perfectly able to discriminate.3. Arguments are given for studying the behaviour of the parasite in order to answer the question of host discrimination. Five examples of parasites are presented that would erroneously have been classified as non‐discriminators on the basis of the first definition, since they all superparasitize.4. It is suggested that the meaning of the term ‘host discrimination’ be extended to include the ability of a parasite to distinguish hosts with different