The km22 RS Energy Calibration using the RSMON corrections improves 
the results greatly on a run-by-run basis.  However, there are
a few phototubes that seem to get worse or give not what we 
want/expect/wish/pray for.  What we want is to have the gain
vs run as flat as possible.

I have created 9 sets of gain files based on the following runs ranges:
49965:50013, 50014:50029, 50030:50057, 50058:50080, 50081:50120,
50121:50142, 50143:50172, 50173:50191, 50192:50218.

To see the plots of gain vs run # for each tube look at the
following website:
http://www.phy.bnl.gov/~benjil/meetings/video/200303/
->   2002.ps shows the gains before the RSMON corrections were applied
->   2003.ps shows the gains with the RSMON corrections applied
->   rdrun.ps shows the Kentaro's run-by-run factors vs. run #



**These tubes were much worse on a run-by-run basis with the 
**RSMON correction

L2 S16 E1:
   runs 50121-50142: gain is low
   runs 50173-50218: gain is high

L8 S22 E2:
   runs 50030:50057: gain is high
   runs 50143:50172: gain is low
   runs 50030:50057 - 50143:50172 there is an overall 
    constant downward trend.


**These tubes were improved on a run-by-run basis with the 
**RSMON correction

L1 S1 E2: erratic

L1 S4 E2: 
   runs 50030:50057 - 50058:50080 there is a large jump in gain.

L1 S7 E1: erratic

L1 S10 E1: 
   runs 50058:50080: high
   runs 50121:50142: low

L1 S13 E2:
   runs 50058:50080: much to high
   runs 50081:50120: high
   runs 50058:50080 - 50173:50191 there is an approx. constant
    downward trend

L1 S19 E2:
   runs 50121:50142,50173:50191,50192:50218: the gains are
    about constant for these 3 blocks of runs and the value of 
    the gains for the other 6 blocks are of a different constant.

L3 S4 E2:
   The first 6 blocks of runs are about constant with a gain of ~18. 
   And the last 3 blocks are all ~16.  So there is a discreet
   change between 50121:50142 - 50143:50172.