The effect of scopolamine on remembering was examined in a delayed conditional discrimination procedure with rats. Remembering was quantified by a negative exponential function fitted to estimates of discriminability derived from a signal detection type of analysis. This function had two parameters: a measure of initial discriminability of the sample stimuli in the absence of a memory requirement (at zero delay) and a measure of rate of forgetting. Eight rats were trained on an auditory delayed conditional discrimination task until they were showing stable performance. Each rat then received doses of 0, 0.005, 0.014, 0.042, 0.125 and 0.375 mg/kg scopolamine IP in a saline vehicle. There was a highly significant, largely linear, decrease in initial discriminability. This was obvious even at the lowest dose of scopolamine. Poorer memory, as demonstrated by an increase in b, was only apparent at the highest dose. Significant changes in per cent of correct responses were also only obtained at higher doses. These results show that initial discriminability and rate of forgetting are pharmacologically as well as theoretically independent. They suggest that the measure of initial discriminability used here is a particularly sensitive measure of at least some types of cholinergic dysfunction; and they also suggest that effects of scopolamine in other working memory tasks could be more a result of changed stimulus processing than of impairment of memorial processes.