Unlike a motion tween, a shape tween requires that you start and end with a native Flash shape, rather than a symbol. Flash uses the vector information of the shape (or shapes) in the start keyframe and the changed vector information of the shape in the end keyframe to construct intermediate frames.
