On the USNO-Flagstaff pixel-sever, the red/blue magnitudes choice is simply for the screen plot that comes up, _not_ for the image data. At the moment, you get all available scans, sometimes 8-10 images in several colors. Unless a star has an extremely red color, image brightness compared to nearby field stars should be essentially the same on the POSS-I O plates (i.e. the blue-light exposure) as on the POSS-II IIIa-J images since the passband is nearly the same (slight extra red-end sensitivity on the IIIa-J plates, but not much). In the south, the POSS-I O (if covered) should be the same as the ones with prefixes 'ao' or 'sb'. If you do see change in relative brightness amongst all the blue- sensitive plates, then you immediately suspect variability. \Brian