There is growing evidence that targeted instruction can improve diagram comprehension, yet one of the skills identified in the diagram comprehension literature--coordinating multiple representations--has rarely been directly taught to students and tested as a classroom intervention. We created a Coordinating Multiple Representation (CMR) intervention that was an addition to an intervention focused on Conventions of Diagrams (COD) and tested their joint effects on diagram comprehension for near transfer (uninstructed biology diagrams), far transfer (uninstructed geology diagrams), and content learning (biology knowledge). The comparison group received instruction using a previously validated intervention that focused exclusively on COD. Participants were 9th-10th grade biology students (N?=?158 from two schools), whose classes were randomly assigned to COD alone or COD?+?CMR conditions and studied with a pretest-posttest experimental design. Both groups showed significant growth in biology knowledge (d?=?0.30--0.53, for COD and COD?+?CMR, respectively) and biology diagram comprehension (d?=?0.28--0.57). Neither group showed far transfer. Analyses of student work products during the interventions suggest that gains were not simply due to the passage of time, because student effort was correlated with gains in both treatment groups. Directions for improving future CMR interventions are discussed.