Strong correlational evidence suggests that involvement in the arts improves students' academic outcomes and memory of learning events [1-3]. It is unclear whether the improved outcomes are the result of general exposure to the arts, the integration of arts into content instruction, the use of effective instructional practices, or a combination of these factors. Moreover, as a growing number of studies suggest that arts-integrated pedagogy enhances learning, few empirical studies have explicitly examined the direct effect of an arts-integrated curriculum on learning and specifically on students' memory for non-arts academic content. Thus, this study sought to determine the effects of arts-integrated lessons on long-term memory for science content. We hypothesized that embedding arts-based activities into conventionally taught lessons would produce learning outcomes as good as or better than traditional instruction. This paper describes the results of a randomized control trial that measured retention of science content using arts-integrated science units and matched units employing convention science instruction. The study was conducted in 16 fifth-grade classrooms in an urban mid-Atlantic school district.