This study investigates whether Virtual Manipulatives (VM) within a Physical Manipulatives (PM)-oriented curriculum affect conceptual understanding of electric circuits and related experimentation processes. A pre-post comparison study randomly assigned 194 undergraduates in an introductory physics course to one of five conditions: three experimental conditions with different PM and VM sequences and two control conditions with only PM or VM. Conceptual tests assessed students' understanding. Instructors' journals, video data, and interviews provided process-related data. Results showed interplay between manipulative and circuit types. For simple circuits, PM and VM use similarly impacted students' understanding. However, VM better facilitated understanding than PM for complex circuits: PM users, unlike VM users, encountered process-related problems that prevented development of an appropriate conceptual model because only VM afforded a view of current-flow. When students used VM before PM for complex circuits, they developed the appropriate conceptual model to use in the PM phase.