Jonathan, When in arbitrary precision mode, rotating the image by exactly +-90 degrees causes the rotation angle in the TAB-window to be shown as (-)57.296 degrees (image calculation, however, remains correct at +-90 degrees). Saving the current settings as a par gets the (-)57.296 degrees rotation to be written to the file (that's how I discovered this behaviour). Changing the par's "center-mag=" by hand to (-)90 degrees and loading it into Fractint results in the same (correct calculation, but wrong TAB display), but now use of any other user interface screen (i.e. "x") causes Fractint to continue with the (-)57.296 degree angle. An angle close to +-90 degrees (not getting rounded to 90) seems ok (the nearest I managed in 1024x768 was -89.995 degrees). Rotating the image to +-90 degrees while still in floating point and *then* zooming in keeps the TAB display at the correct rotation angle, but now Fractint refuses to switch to arbitrary precision. This is DOS-Fractint 20.04p9 in MS-DOS 6.22, Fractint 19.6 behaves likewise. The original par ("longtime.par" by Noel Giffin) has a "reset=1921" entry, so this seems to go back quite a while. It certainly isn't a catastrophic malfunction, but I thought you should know. Regards, Gerald