The table below contains links to plots showing the chi-squared values returned by PROPHOT profile-fitting as a function of magnitude (with no active deblending) from three early northern nights when image elongation due to focus errors was severe. The extent and duration of the elongation can be seen in the top panel of the image quality plots for each night also linked in the table below. Those plots show the mean image aspect ratio for each scan.

The image quality was uniformly poor on the night of 970610n. On 970614n and 970712n, the image quality varied during the night. The respective chi-squared distribution plots show that for bright (but non-saturated) sources, the chi-squared values rise systemmatically with increasing brightness. The trends are most severe on 970610n. The scatter in the chi-squared values is larger on 970614n and 970712n, which is consistent with the degree of elongation variations during the night.

The horizontal lines at chi-squared = 1.8 denote the current threshold above which PROPHOT would attempt active deblending. Active deblending would attempt to operate on a large fraction of the high SNR sources on these nights, and would "succeed" in deblending many single sources because the elongated profiles are better fit with multiple point-spread-functions. The chi-squared limit for active deblending would have to be set so high to avoid spurious deblends that it is effectively useless. Active deblending should be turned off all together for these nights.

At this time we do not have a measure of mean chi-squared values for high SNR sources as a function of mean image elongation. Selectively turing off active deblending scan-by-scan based on mean image elongation is not currently possible.