RESUMO
Pigeons were maintained on a fixed ratio (FR 9) schedule of reinforcement for correct matching-to-sample responses. Included in the test situation was a vertical array of lights, illuminated in relation to the successive steps of the fixed ratio. All five subjects showed regular decrements in matching errors across the sequence of unreinforced responses within the ratio cycle. In the form of a randomly introduced probe, the stimulus situation (array of lights) appropriate to having seven of the FR 9 steps already completed was occasionally introduced at the beginning of an FR cycle. Reinforcement followed the illumination of the two remaining lights by two correct matches. The number of errors in this probe condition was sharply lower than the errors characteristic of the first two steps of the basic FR 9.
