Problem(s) Encountered:

At one point in the evening (0145 local/0745 UT) we found a guide star and set up the autoguider, but within 30 seconds the guider started moving the telescope by 10-20 arcseconds at a time, even though we could see the guide star clearly in the guide camera box (briefly, until it lost itself). It basically moved in larger and larger circles until we stopped auto-guiding. We tried changing guide stars, and the problem repeated. We guided by hand for that target, but didn''t encounter the problem again.

Solutions:

This sounded to me like the rotation angle on the guider software was not set correctly for a nonzero rotator angle.  Verified with the observer that this was indeed the case.  [-90 - rotator position]