Analysis of Un-De-Dithered Focus Plates

I've performed a simple analysis of the width of the trail in the un-de-dithered HRC-I focus plate images. The analysis is as follows

the exposures were divided into one second time slices, and the image center of each slice determined using a sigma-clipping algorithm, with a 1 sigma clip radius.
the distance from each event to the center for its time slice was determined
distances which were greater than 20 pixels from the derived center were tossed, to get rid of background events.
the distances were then grouped into 15 second slices. slices with less than 20 events were tossed. the mean, RMS, and 80% half width of the remaining slices was determined.

Spot diagrams and the statistics are plotted below. The center of the time slices is plotted on the spot diagram.