Extensive anatomical and psychophysical data suggest that macaques can serve as surrogates for man in studies of the neural substrates of visual mnemonic processing. However, the extent to which mnemonic-mechanisms in macaques can be validly compared to those in man depends critically on the demonstration that the cognitive strategy and temporal dynamics of mnemonic retrieval are congruent For the two species. Seven human and six macaque subjects were tested on identical versions of an item recognition task that required the classification of probe stimuli as positive or negative according to whether or not they were members of a previously defined set of target stimuli. For the human subjects, reaction time increased by an average of 24 msec/target. The macaques were able to respond as accurately as the human subjects, but each additional target resulted in a reaction time increment of only 7 msec. A detailed analysis of the statistical properties of the reaction-time distributions indicated that these data do not reflect a between-species difference in the efficiency of execution of otherwise comparable, serial-exhaustive, retrieval mechanisms. Rather, the data suggest that human subjects engage two memory-load dependent processes before generating a response, while macaques execute only one of these. Additional data indicate that a reaction-time approach toward the analysis of the stages of mnemonic processing can provide new insights into the specific nature of mnemonic deficits induced by brain damage.