TIME HORIZON AND CHOICE BY PIGEONS IN A PREY-SELECTION TASK

In two operant simulations of prey selection, the choice behavior of pigeons and the predictions of a model of the effect of session length on optimal prey selection were compared. All but one of the predictions were confirmed. In the first experiment, a higher proportion of "bad" items (comparatively long delays to food) was accepted in 10-min sessions in one operant chamber than in the first 10 min of 20-min sessions in a different chamber. The proportion of bad items accepted increased toward the end of the 20-min sessions. In the second experiment, the birds were exposed both to 20-min sessions in one distinctively marked operant chamber and to sessions of variable length, with a mean of 20 min, in another. A higher proportion of bad items was accepted in the variable-session chamber than in the fixed-session chamber at the beginning of the 20-min sessions. This difference decreased toward the end of the sessions but did not reverse as expected. In both experiments, an independent test analogous to the peak procedure was administered to establish whether the birds discriminated among the different session lengths.