As second language (L2) research adapts to the dynamics of multimodal instruction, researchers are exploring the role of captioning and textual enhancement (TE) on learner development. The present experiment assesses the differential effects of textually enhanced captions on learners' elicited imitation of L2 Spanish grammar. Twenty-eight English-speaking intermediate L2 Spanish learners saw 3 videos that focused on "gustar"-type verbs, the preterite--imperfect contrast, or the subjunctive. Each video included 3 conditions: sentences without captions (NC); sentences with target verbs highlighted entirely (TE1); or sentences with highlighting on inflectional morphemes (TE2). Using a novel written elicited imitation task, we uncovered consistent positive effects of both TE conditions compared to NC, with an added advantage of TE2 for the subjunctive. The findings demonstrate that (a) TE with captions contributes to increased L2 accuracy, (b) the elicited imitation of some grammatical structures is more conducive to enhancement than others, and (c) there is space for future investigation into the factors that mediate the effectiveness of multimodal interventions.