know of none using a serial probe recognition procedure. One recall experiment ( Postman & Riley, 1959) is distinctive in that it used a serial learning procedure from which serial position functions were obtained. They showed a rebound of the primacy effect much as we have shown here. Most of the studies used a paired-associate design from which serial position functions are not usually obtained.
We feel compelled to make some mention of the role of rehearsal before leaving this discussion of the probe-delay effects, because rehearsal is currently thought to be responsible for the primacy effect ( Waugh & Norman, 1965; Atkinson & Shiffrin, 1968). Longer probe delays provide more opportunity for rehearsal, but rehearsal during the probe delay probably cannot account for rebound of the primacy effect. Consider the following: If the subjects can retrieve the items in order to rehearse them, then why can't this same retrieval be instrumental in the correct choice response at short as well as long delays? Indeed, if rehearsal functions to bridge the delay, then primacy effect should be better at short than long delays because of occasional rehearsal failures.
Pigeons acquired serial probe recognition performance with list lengths as long as four items and exhibited primacy as well as recency effects in their serial position functions. The primacy effect figures prominently in many memory theories, and from a comparative memory stand-point it is important that pigeons, like humans and monkeys, exhibited this effect Pigeons have difficulty performing with variable list lengths and so unless this difficulty can be overcome, their memory-scanning processes cannot be tested and compared to monkeys and humans. Unlike monkeys, however, they do not readily develop short-cuts to the memory task and, for example, memorize the correct responses to the probe items.
Monkeys (and humans too) when given the opportunity, memorize the correct responses to probe items and thereby short-circuit the memory process of comparing the probe item to their memory of the list items. In order to thwart their probe memorization strategy, items were changed on a regular basis before they had a chance to memorize them. The "U" shaped serial position functions and the systematic changes in the serial position function with probe delay demonstrated that all three species were making a memory comparison.
The probe-delay effects reflect what may be an important mechanism of memory processing generally: an absence of a primacy effect at short probe delays, its emergence at intermediate delays, its dominance of the serial position function at long probe delays. These probe-delay effects are demonstrated and reported here for three different organisms: humans, monkeys and pigeons. These results may be important for the human memory literature from the standpoint of development of the primacy effect and may support interference theory and the release from RI. Such results, dealing with the subtle and fine-grained inner workings of the memory system, have heretofore been confined to human memory research. It is of paramount importance that these results have been