THE SUB-PLATE METHOD AND WIDE-FIELD ASTROGRAPH PLATES

have individually rereduced the 64 11 degrees x11 degrees astrograph plates, which comprise the ''South Polar Cap'' region of the Yale zone catalogs [Hoffleit, Trans. Yale Univ. Obs. 31 (1971)], by both the sub-plate algorithm and the traditional global plate model approach. The histogram of positional errors for the comparison stars is significantly narrower for the sub-plate method than it is for either the original plate model used on this material or a more scientifically justifiable one. The amount of this improvement is 22% as measured by the average standard deviation of the 4193 comparison star positional deviations. The comparison stars, whose positions were taken from the IRS catalog, were not included as reference stars in any of the reduction processes but were used as an independent, external check on the plate solutions. We show that the reason the sub-plate technique is superior is that these wide-field astrograph plates suffer from systematic effects that the global plate models cannot successfully represent. The amplitude of the positional errors from the sub-plate method matches what one would predict based on the measuring error of these plates (i.e., 0.'' 4 per coordinate). An Appendix presents the theoretical framework for weighting the different positional estimates obtained by overlapping sub-plates. (C) 1995 American Astronomical Society.