craigd, you have a valid point. The bar with the ball bearings is 36", roughly 3/4" is resting on each pillar, so now you have 34 1/2" and at half or 17 1/4" would be I would think, the most sag. My theory, since the balls are closer to each pillar, the barrels weight is being distributed evenly across these two points and does not effect the center at all.
Most of the weight would be on the left side, naturally where the barrels are joined and where the lug is, which is forward of where the first ball is.
In the scenerio you suggested, it would be the same, once you moved the barrel to read that ball under it, then zeored it and moved the barrels over it again, the readings would be the same as if the barrels weren't moved at all.
In my post before this, I made a mistake, the aluminum block holding the indicator was filed square to fit the square tube, but it came out loose, and as Gnoman stated, it has to have clearance for it to slide. But in re-checking it more than once, it comes back to within that .001-.002 from the original.
The pillar blocks were drilled round and a set screw holds it tight to each pillar.
