When working with intercube rules, as in the case of virtual cubes, rule statements reside in the target (the virtual) cube, while feeder statements always reside in the source cube.

It may be easier to understand if you think of your feeders pushing and pulling. Feeders do the pushing of data from the source cube (so, they will reside in or be associated with the source cube) while rules do the pulling of data into the destination cube (so, they will reside in or be associated with the destination cube). This is true for standard TM1 cubes as well as virtual TM1 cubes.

For example, if the virtual cube references data in the standard cube, the TM1 rules to reference the data in the standard cube will be associated with the virtual cube. The feeders to support these rules will reside in the standard cube: