The distribution is particles bigger than or particles smaller than. Whoever worked out the equations for the original model was probably taught the "other way" to yourself so did it that way. For a cascade impactor, if I weigh each tray I get a mass, logically I see zero powder on the tray with the largest holes, so 100% of mass is under that size. I don't know the size of the smallest particles as they get through all of the trays, so any mass missing must be under the size of the holes in the smallest tray.