Following up on a discussion yesterday, and with apologies to those of you who know this already, here is a plot of the signal-to-noise as a function of aperture size, assuming Gaussian seeing. The x-axis is in units of the FWHM seeing, but note it is aperture radius, rather than the more usual aperture diameter. You can treat the y-axis as arbitary units, but if you want the full details of the peculiar units see the paper at: http://ukads.nottingham.ac.uk/cgi-bin/nph-bib_query?bibcode=1998MNRAS.296..339N&db_key=AST&data_type=HTML&format=&high=4224ae18ba15556 As the aperture gets smaller it allows for the light loss, but also the decrease in noise because there is less sky.

And the bottom line. Well, Richard used an aperture diameter of 2 arcsec. If the seeing was really 0.8 arcsec, then the aperture radius is 1.25 FWHM, so the the best possible improvement in signal-to-noise by using a smaller aperture would be 0.6/0.45=1.33. Useful, but not huge (bearing in mind the uncertainties).