A classroom study was designed to test the hypothesis that explicit knowledge is used by second-language (L2) learners in a free written response task if that knowledge is present. Eighty-one 12–18-year-old learners of Dutch as an L2 took part in a computer-assisted language learning experiment receiving either explicit or implicit instruction about two grammar structures. The ability to use these structures was measured at three points in time by means of an untimed grammaticality judgment task and a free written response task. Explicit and implicit instruction promoted the use of the target structures in free response tasks equally effectively. However, for one structure, both facilitative and inhibitory effects of explicit instruction were observed if first language similarity was taken into consideration.