Monolingual studies contrasting memory for positive versus negative emotion-laden words have generally used single-trial paradigms and have produced inconsistent results (no difference or an advantage for either positive or negative valence). However, monolingual studies with multiple presentations of stimuli have consistently found a positivity advantage in recall. No bilingual study has examined whether L2 testing, using a multi-trial procedure, will also produce a positivity advantage. We report two experiments in which L1 and L2 participants performed three learning trials (aural exposure, oral recall), followed by multiple delayed oral recall trials and a recognition trial, using lists of English words from ad-hoc semantic categories, with equal numbers of positive versus negative valence words. Results, including an overall positivity advantage, a greater positivity advantage in L2 than L1, and greater valence-based clustering in L2 than L1, were discussed in terms of the effects of stimulus exposure and gist consolidation.