This is what I know about SAS procs. Why don't you use a loess smooth? If you really want a third. degree polynomial I know about interpolation but don't use that..I suppose you could use some kind of third degree polynomial regression depending on your data. Good luck, David Booth