The present Aptitude-Treatment-Interaction (ATI) study investigates the learner characteristic spatial ability (aptitude) and the variation of mental-animation prompts (treatment: no vs. mental-animation prompts). A group of high-school students (N = 94) learned about a biology topic through learner-paced multimedia instruction. Some of the learners received mental-animation prompts and others learned without prompts. A fine-grained analysis with spatial ability as continuous aptitude variable and mental animation as treatment showed a positive learning effect of animation prompts in learning outcomes of processes, but not in knowledge about structures. In addition, spatial ability only modified the relationship between animation prompts and learning when analyzing knowledge about processes. Specifically, only learners of low to medium spatial ability profited from the prompts while learners with very low or high spatial ability had comparable results when learning with or without prompts. In addition, only learners with high spatial ability rated their cognitive load to be significantly higher when learning with prompts. Results align with the assumptions of the production deficiency of learners with low to medium spatial ability, mediation deficiency of learners with very low spatial ability and stable learning performance of learners with high spatial ability whatever the learning situation offers.