Got the final SE corner(3) L4-C leveled this morning. Attached is a plot of spectra with five before and three after leveling.  H1 & H2 show a poor (low) response at lower frequencies before leveling.  H4 has good response before and after and H3 has not been leveled for either trace in this plot.  I did check the sprectrum fter leveling and H3 still looked fine.  I does beg the question if it is possible to level an L4-C out of goodness...  Photo sequence attached shows accessing the L4-C mounting screws (after removing the two Actuator adapters), the shims under one corner of the L4-C Mount inside the Crossbeam Foot and before and after levels.  I know I know, it sure doesn't look like it is level.  But, we have no mechanism to roll the mount to center the bubble, and; leveling the L4-C along its long axis is what matters.  The second ring of the level is about 3mrad.  The L4-Cs that were performing poorly, H1 & H2 did require more shimming than the other two corners.  Not a factor or two but larger then 50%.  The effective shimming in the worse case was 18 mrad were as in the least case it was 11 mrad.  Shims are used on just one of the two lower legs of the three point mount, the other leg is not accessible.  So the shim effect is half the shim thickness.  Shims used here were from .086 to 0.134".