A note on choice of end members in representing certain systems and on a possible alternative to Vegard's rule