-- Description for the plots --

Only the runs with flag-0 in CNI polarimeter spread sheet are used.

Specific luminosity for each bunch at CNI polarimeter is calculated;
the relative sizes are considered as the measure of the bunch quality.
(beam positions have been slightly changed bunch by bunch?) 
The definition is 

 Specific luminosity = Total counts in CNI polarimeter/Wall current monitor

There are four plots in each measurement,
(Upper left)
 The specific luminosity distributions. There are two bunches (1st, 11th) 
 known to be anomalous and should be thrown out. In addition, two groups of
 problematic 5-bunches. The location is fixed to be 16th-20th and 36th-40th   
 and considered to have something to do with collisions (corresponding to 
 the collision with abort gaps at some interaction points).
 Corresponding bar charts are colored as green.
 The bunches whose fit chi^2/ndf exceed 10 are plotted with black, and the
 red triangle marker is plotted above it.

(Upper right)
 Horizontal projection of the upper left plot. Anomalous bunches (1st,11th,
 16-20th and 36-40th) are plotted with green again.

(Lower left)
 Total counts at CNI polarimeter is plotted as a function of the bunch id.

(Lower right) 
 Wall current monitor information. 

It should be noted that sometimes the wall current monitor information is 
collapsed and shows a avalanche shape structure (for instance, see run 856)
or sometimes no WCM info available (see run 799). This ends up with the 
funny structure in the specific luminosity distribution. The phenomena is 
seen in both rings, and indicates that the automatic bunch selection from 
certain criteria would fail.

As a result, it can be said that the specific luminosity distribution will 
be a good confirmation tool for the bunch selections, but cannot be used as 
a condition itself.