Jim, thanks for the help: that seems to pin it down to a problem with the data, and shows that it's not just me (always comforting). Analysts, is there any chance that you could dig out some datasets within a few days of AP94 for me and follow the procedure Jim suggests? Thanks, Martin