We have NM symposium today and another symposium tomorrow. I tested this by doing my data with the new method and old programs and results were identical including failures except that weights do change. If you want less change even with old methods set DOFLAG=-6 or some such. I will look but it will take time. Eric