Six short-term memory tests from Guilford's structure-of-intellect model were administered to 60 subjects. The subjects were then presented with a four-category concept problem in two phases. During the first 36 trials, randomly selected instances were repeated in order to determine subjects' ability to recall previously presented instances. Beginning with the thirty-seventh instance, the previous instance was made available for 30 randomly selected subjects, while the other 30 subjects were not provided with an available instance. Factor analysis of the six tests yielded two factors: Rote Memory and Classes Memory. The ability to recall past instances was significantly correlated with Rote Memory. Presence of an available instance did not significantly facilitate mean performance, but it did significantly alter the relationship between Rote Memory and performance in the concept task.Recently, investigations concerned with the role of various intellectual abilities in different concept-learning tasks (e.g.