Choice biases in delayedmatching-to-sample duration with pigeons: Manipulations of ITI and delay illumination

Two experiments investigated the effects of similarity between intertrial interval (ITI) and delay illumation on the choose-short effect. Different groups of pigeons learned to match ` s̀hort’ ’ (2 s) and ``long’ ’ (6 or 8 s) food samples to green and red test stimuli in a matching-to-sample procedure with a 5-s training delay. Subsequent 10- and 20-s delay tests revealed choose-short effects if the ITI and delay were both illuminated (i.e., group ON±ON), if the ITI anddelay were bothdark (i.e., groupOFF±OFF), and if the ITI was illuminated and the delay was dark (i.e., group ON±OFF). In addition, either a choose-short effect or a choose-long effect was observed if the ITI was dark and the delay was illuminated (i.e., groupOFF± ON). Results are incompatible with the confusion/instructional failure viewof the choose-short effect. The delayed matching-to-sample (DMTS) task with duration samples has been a parti-cularly useful procedure for examination of codingprocesses inpigeons ’ workingmemory (for a review, see Grant, Spetch, &Kelly, 1997). Each trial begins with a temporal sample, consisting of either a ` ̀ short’ ’ (e.g., 2-s) or ``long’ ’ (e.g., 8-s) duration of a stimulus event (e.g., presentation of a food hopper), followed by an opportunity to choose between tw