To assess the effects of methylphenidate on working memory, pigeons were trained in a delayed matching-to-sample task. Delay interval duration (0.2, 1, 3, 6, or 12 sec) was varied within sessions in order to separate delay-dependent from delay-independent effects of the drug on performance. A reduction in the sample response requirement from five responses to one response effectively reduced attention to the stimulus and impaired overall accuracy. Methylphenidate was administered in doses of 0.0 (saline control), 0.25, 2.5, and 10 mg/kg. Relative to performance with saline, accuracy was significantly reduced with 10 mg/kg methylphenidate to the same extent in both fixed ratio (FR) 1 and FR 5 conditions. The smaller doses had no effect, and there was no evidence that accuracy improved with drug administration. Intercepts and slopes of exponential functions fitted to measures of discriminability plotted as a function of delay showed that methylphenidate affected delay-independent aspects of performance (initial discriminability), but not delay-dependent aspects (rate of forgetting).