In this paper we describe a new memory test in rats, based on the differential exploration of familiar and new objects. In a first trial (T1), rats are exposed to one or to two identical objects (samples) and in a second trial, to two dissimilar objects, a familiar (the sample) and a new one. For short intertrial intervals (approximately 1 min), most rats discriminate between the two objects in T2: they spend more time in exploring the new object than the familiar one. This test has several interesting characteristics: (1) it is similar to visual recognition tests widely used in subhuman primates, this allows interspecies comparisons; (2) it is entirely based on the spontaneous behavior of rats and can be considered as a 'pure' working-memory test completely free of reference memory component; (3) it does not involve primary reinforcement such as food or electric shocks, this makes it comparable to memory tests currently used in man.