>> Yes, but Yamaoka-san was talking about the CV itself, not >> the sequence stars (which are the x,y guys). Yep, I figured that out about 2 sec after hitting the 'send' button--- duh! >> I doubt B1.0 can give RA,DEC to 35mas as you indicated ... Well, that's what the catalogue says. Presumably that's the mean-error on the five plates, so in principle about 80mas per axis per plate. If the star is isolated and well-placed on the plate, then the mean could certainly come out pretty small, though that would be distinctly smaller error than for a large sample. \Brian