why are astrometric residuals smaller in I than V?

[Stupendous Man <richmond@a188-l009.rit.edu>]

  Andrew wrote:

>   Why the long exposure I-band images are consistently
> better than the V-band when the PSF is clearly fuzzier
> for the I-band, I do not know

  If you're down at the 0.2 arcsec level, it _could_ be 
differential refraction, of two sorts:

     a) stars near the "top" of the frame will be refracted
        by less than stars at the "bottom".  Check for this 
        by plotting residual versus altitude, or zenith angle,
        or some quantity which correlates with these

     b) refraction by the atmosphere is larger in the blue than the red,
        and it depends slightly on the color of each star -- blue stars
        will appear to be refracted very slightly more than red stars.
        Moreover, the DIFFERENCE between the refraction of blue stars and 
        the refraction of red stars will be larger in blue passbands than
        red ones.  Check for this by plotting residual versus
        (B-V) color, which is available in the Tycho catalog, or versus
        (V-I) instrumental magnitude, from the Mark IV frames themselves

                                            Michael Richmond