Thanks very much for all this, Russell. You have added value, as they say, to my article. (I'm not understanding "must be over a range in the underlying data," though.)