Children's performances using three different mouse interfaces to control point-of-view (POV) navigation in a prototype of a CD-ROM based "virtual forest" were assessed. Results indicate that while children readily understood POV movement and were able to use all three interfaces successfully, each interface was less than optimal for different reasons. An assessment of the strengths and weaknesses of each interface in light of the intended usage scenario was conducted, and the least problematic of the three was selected for the system. KEYWORDS Children; interface; virtual travel