This study extends previous input-based grammar instruction research (for reviews, DeKeyser and Prieto Botana 2015; Shintani 2015) by comparing two types of input-based practice, each with the same explicit information, for learning L2 German definite article case-marking cues (der, den). Participants (N=138, aged nine to 11) received explicit information followed by either task-essential practice in making form-meaning connections (referential activities from Processing Instruction) OR task-essential practice in spotting the form (noticing activities). Both interventions yielded equivalent durable gains across six ecologically valid tests of comprehension and production (written and oral modalities), compared to negligible gains in a Control group. The findings revealed that, following explicit information, input practice requiring noticing of the target feature (as proposed by Svalberg 2012) was equally effective as task-essential form-meaning connection practice, shedding important light on previous claims in the research agenda on task-essential input practice (e.g. Marsden 2006; Marsden and Chen 2011). Responding to calls for ecologically valid effect-of-instruction research (Mitchell 2000; Spada 2015), this classroom study demonstrates the efficacy of grammar practice for young learners within input-poor foreign language classrooms.