This study investigated the effects of negative feedback type (i.e., explicit correction vs. recasts), communication mode (i.e., face-to-face communication vs. synchronous computer-mediated communication), and target structure salience (i.e., salient vs. nonsalient) on the acquisition of two Turkish morphemes. Forty-eight native speakers of English with no Turkish background carried out two communicative tasks during which their errors on the target structures were treated according to their feedback group. Oral production, comprehension, and recognition tests were used to measure learners' resulting performance. A clear advantage was found for explicit correction over recasts in the oral production and comprehension tasks on both immediate and delayed posttests. Results also showed that neither communication mode nor target structure salience moderated the difference between the negative feedback types but both factors made independent contributions to feedback effectiveness. (Contains 5 notes, 4 tables and 1 figure.)