While strategy instruction research generally focuses on the effect of the teaching on learners' use of the strategies targeted for instruction, the present study examines the "wash over" effect on learners' use of pre-existing, non-target strategies. The study involved a treatment class and a comparison class in the ESL oral classroom in Hong Kong. A quasi-experimental design was adopted in which the treatment class received metacognitive strategy instruction to enhance students' ability to manage English group discussion tasks. A data collection method comprising observations and stimulated recall interviews that aimed to tap respectively overt strategy use and covert strategic thoughts was employed. The findings indicate that, for the treatment class, while there seemed to be no "wash over" effect from target to non-target strategies in the frequency of use of the whole sample of non-target strategies over time, there were increases in the frequency of use of individual, non-target strategies that were pertinent to the specific nature of small group discussion tasks. The findings are discussed with respect to the differential and desirable effects of the strategy teaching and to the relationship between task type and strategy choice. Implications for strategy instruction will also be drawn. (Contains 4 tables.)