Abstract
Fifty-two Ss solved six conjunctive concept attainment problems. All problems had 2 relevant and 5 irrelevant attributes which were presented in a 6 by 6 random replicated Latin square design. All stimulus cards were continuously in S's view. Each stimulus card choice of S was recorded. From the protocols of S's card choices a record was made of the rate at which information about the concept became available. It was found that S's potential acquisition of information was efficient early in concept attainment but decreased in efficiency as the amount of information to be retained about the concept increased.