into 20-trial blocks. Selection of list length, list items, same/different trials and serial position were determined in a pseudorandom fashion. Within a 20-trial block, the number of trials at each list length and the number of same/different trials occurred with equal frequency. List items were drawn without replacement from the 140 item pool of pictures every 20 trials. On Same trials, the serial position of the target (match to probe) was sampled with equal frequency in a pseudorandom fashion for each list length. On Different trials, unused slides within the 20-trial block were designated as probe items. The 1,200-trial sequence was divided into five smaller sequences of 240 trials for daily experimental sessions. Both monkeys and humans were given one daily session with one of the five 240-trial sequences.
It is important to note that the stimulus sequence was sufficiently large to ensure that each session presented a unique stimulus sequence. This made it impossible for subjects to learn the order of stimulus presentation or to form stimulus-response associations. Previous studies (see Chapter 20, by Wright, Santiago, Sands, & Urcuioli, this volume) have shown that if such precautions are not taken, monkeys are quite capable of memorizing correct responses to probe items.
A trial began with a 100 Hz "ready" tone that signaled the availability of a trial to the monkey. A downward press of the lever terminated the "ready" tone, and after a 0.5-sec delay, a list containing either 2, 4, 6, 8, or 10 items was presented on the top screen. Each item in the list was presented briefly for 80 msec. The major variable of interest in this experiment was the interstimulus interval between successive items. Interstimulus intervals of 80 and 1000 msec were selected on the basis of previous data on pictorial recognition ( Intraub, 1980). One sac following the presentation of the last item of the list, a probe item was displayed on the bottom screen and remained in view until the monkey responded or until a 2-sec response cutoff had elapsed. Upon presentation of the probe item, the subjects then responded either Same or Different by moving the response lever to the right or to the left, respectively. Correct responses were followed by a short tons (3000 Hz. 1-sec duration), a reinforcer and a 4-sec intertrial interval. Incorrect responses or failures to respond within 2 sec were followed by Illumination of the houselight, omission of the reinforcer, and a 5-sec intertrial interval. Failures to respond within the cutoff interval were treated as incorrect responses. All human subjects received five sessions using a similar procedure at each rate of presentation. Oscar and Felix received 19 and 17 experimental sessions at each presentation rate, respectively.
The humans were tested under conditions similar to those of the monkeys except that they set in a chair 100 cm from the stimulus panel and held the response lever in their right hand.