This paper discusses the effectiveness of using two task-types, dictogloss and opinion-gap tasks, in focus-on-form lessons to elicit language-related episodes (LREs) and produce modified output. The participants were 40 Form 6 students in a Malaysian secondary school in two intact classes. One group was taught using opinion-gap tasks and the other using dictogloss tasks. Results show that almost double the number of LREs was elicited in the dictogloss group compared to the opinion-gap group. However, more than 50% of the LREs in both groups were unresolved or wrongly resolved. These results imply the need to train learners to notice linguistic errors and engage in negotiations of form and meaning and could also indicate the need for teacher feedback on language use, perhaps during pre-task and post-task activities. The types of linguistic errors learners focused on might be an indication of learner developmental readiness which could influence task and text selection.With the current inclination in second language acquisition (SLA) practices toward task-based language teaching (TBLT), there is a growing concern that the importance of language accuracy will be undermined. Skehan and Foster (2001), for example, argue that a task-based approach, if inappropriately handled, could result in an overemphasis on task outcomes at the expense of improving target language ability. As a result, it could encourage premature fossilization rather than promote interlanguage development. Thus, SLA researchers (e.g., Ellis, 2001;Ellis, Basturkmen, & Loewen, 2001;Loewen, 2004;Leeman, 2007) have argued for the need to incorporate grammar instruction in a meaning-focused lesson. These researchers challenge Krashen's (1985) input hypothesis, which posits that grammar instruction is not an important part of language instruction. Instead, he contends that language is developed through massive comprehensible input of the language. Initially, Long (1985), through the interaction hypothesis, concurred that comprehensible input facilitates acquisition as it provides opportunities to negotiate meaning during a communicative breakdown. Later, Long (1996) extended his hypothesis to include feedback and modified output as important contributors to interlanguage development. He explains that the feedback learners receive on their language production when they attempt to communicate could contribute to language acquisition, as this would push learners to reformulate their productions to make them comprehensible.