Functional studies of the dorsolateral prefrontal cortex and the inferior parietal lobe of the rhesus monkey have implicated the former in spatial mnemonic function and the latter in visuospatial processing. We used the 14C-2-deoxyglucose (2DG) method to assess the contribution of these cortical regions to the cognitive performance of monkeys on working memory tasks. In these experiments, one group of monkeys (WORK) was trained to perform tasks (delayed spatial alternation, spatial delayed response, or delayed object alternation) that specifically engaged working memory processing. Local cerebral glucose utilization (LCGU) rates in the WORK group was compared with LCGU rates for a second group of monkeys (CONT) tested on one of two tasks (visual pattern discrimination or sensory-motor) that relied upon associative memory. The results showed that in comparison to the CONT group, working memory performance significantly enhanced LCGU by 19% in the principal sulcus region of prefrontal cortex and by 11–20% in regions of the inferior parietal cortex corresponding to areas 7A, 7B, 7IP, and 7M. By contrast, LCGU in the auditory cortex was similar for both groups. In all areas examined, metabolic activation peaked in lower layer III where the majority of associational and callosal neurons lie. Correlation analyses of LCGU and behavioral task parameters indicated that LCGU in the parietal subdivisions was significantly related either to the accuracy of performance or to the number of trials completed on the 2DG test. In contrast, LCGU in the principal sulcus was positively correlated with task difficulty. These findings suggest that the enhancement of LCGU in the principal sulcus was primarily influenced by the mnemonic components of the tasks whereas LCGU in the inferior parietal cortex was influenced by their sensory-motor demands. These are the first results showing concurrent metabolic activation of the prefrontal and parietal cortex in monkeys performing working memory tasks and they support the suggestion that these cortical regions represent two important nodes in a neural network mediating spatial working memory in the monkey (Goldman-Rakic, 1988). Further, the present report reinforces the power of the 2DG method for functional mapping as these areal and laminar results could not be readily appreciated at this resolution in any other methodological context.