The photographic calibrations used here are based on eight plates of two fields, all taken on the night of February 23/24, 1995. These fields are called LP543 and M68 by Guo et al. The measured OFAD coefficients in Table 2 are averages in which equal weight is given to all plates.
The focal lengths measured from the LP543 plates are approximately 3mm longer than those derived from the M68 series while the focal length determined from plates taken in the same color of the same field differed by less than a millimeter. This difference is difficult to explain since the images were taken in uninterrupted sequence during the same night. The air mass at which the LP543 plates were exposed is somewhat higher than for those of M68 so an error may have been made in correcting for atmospheric refraction.
The errors can be put into perspective by observing that the models determined for each plate predict measured image positions on that particular plate with an rms precision of 1.5-2micron. The measured values of the OFAD fluctuate between plates within the same series by approximately 3-4micron of position uncertainty, roughly the same as the difference between the theoretical and experimental models one a correction has been applied to the focal length. The rms position difference caused by the 3mm focal length difference between the LP543 and M68 plates is 14micron . This means that uncertainty in focal length is the principal source of error in our knowledge of the corrector's behavior. The experimental focal lengths in Table 2 are an average of the two plate sets.