[daip] FITS /astrometry question

[Michael Bietenholz]

  I have noticed that AIPS (at least TVLAB, CELGAL and EPOSW) seems to
ignore the rotation on the first axis, and rotate both axes according
to the listed rotation of the 2nd (decl) axis.  Am I correct in assuming
that AIPS (and FITS?) requires the axes to be perpendicular, so that
only 1 rotation angle is required, and that it uses the CROTA2 or
the AIPS-header rotation of the 2nd axis for this rotation?

              thanks,             michael