Measure the polarization of at least one zero-polarization star to establish the system polarization zero. Remember that both the telescope and the A&G + ISIS contribute to the zero-point polarization, so decide in advance how you are going to use the instrument rotator: keeping the slit and instrumental coordinate frame approximately (or even exactly) vertical during your observation, or alternatively with the instrument aligned with the equatorial system. In the latter case, for the telescope's contribution to instrumental polarization, you must allow for relative rotator angle when correcting other observations later. For short observations, the best solution is probably to set the slit just one side of vertical and to let it track during the complete observation, the median slit orientation being vertical (use Fig. 5 for planning this); your coordinate frame then has a fully defined relation to the astronomical frame, while the fact that telescope polarization rotates slightly in the course of the observation does not matter as long as it is small enough.
To `depolarize' a zero-polarization star completely, observe it twice, with a difference of about 90 degrees in parallactic angle (Fig. 5). In the Alt-Az frame, the average of the two observations represents pure instrumental zeropoint (exercise for the reader, including the inherent assumptions). The converse holds for a representation in the equatorial coordinate frame: the average is purely the stellar polarization. A reference to zero-polarization stars is given in Appendix H. Remember the cautions about stability and spectral dependence of stellar polarization.