Transfer of oddity-from-compound samples in the pigeon: Some assembly required.

Experiment 1 trained three pigeons on a six-key oddity-from-sample procedure. At trial onset, one of two compound color-form sample stimuli was presented (a white 'X' on a red background or a white circle on a green background). Every second peck on the sample key lighted another comparison stimulus up to a maximum of four stimuli (red, green, circle and X). A peck on the two keys that presented non-matching color and form stimuli, relative to the compound sample, lighted a reinforcement key. A peck to this key produced grain. Errors darkened the comparison keys and repeated the trial. The birds learned to perform without error during 70-80% of the session trials, and showed transfer of oddity performance to eight novel compound samples. Experiment 2 presented the pigeons with novel combinations of familiar color and form stimuli. Each bird performed well above chance indicating that their behavior was under the control of an oddity concept.