Acquisition and maintenance of matching without a required observing response.

Two experiments demonstrated that matching-to-sample performance was improved when an explicit observing response was required to the sample stimulus. The first experiment demonstrated that fewer training sessions were required to establish matching with than matching without such a response. The second experiment demonstrated the dependence of established matching accuracy upon this observing response and the development in two of three subjects of a new overt observing response even when none was required by the experimenter.