A major purpose of the present studies was to determine the effects of varying the relative frequency, or bias, of stimulus presentation at different locations in a five-choice serial reaction time attention task. Training sessions were conducted in which the stimulus duration was held constant at 2 s and, initially, stimuli were presented with equal probability above each of the levers. During testing sessions, however, stimulus duration was either 0.2 or 2 s, and the frequency of presentation was varied among no bias, a middle bias, a left bias or a right bias condition. The training conditions were then changed such that the frequency of presentation was always a left bias, and test sessions were again conducted with no, middle, left or right bias. In the presence of the 0.2-s, but not the 2.0-s, stimulus, the percentage of correct responding on each of the choice levers during test sessions varied directly with the bias conditions in that the percentage of correct responding was highest at the most frequently presented location and lowest at the least frequently presented location. In addition, scopolamine, but not metscopolamine, increased omissions while also increasing anticipatory responses in rats trained under left-bias conditions and tested under right-bias conditions. The present findings suggest that varying the frequency of stimulus location can provide additional measures of attention in the five-choice serial reaction time task.