Butlersrangers you are on to something there. I have thought of a way to measure actual headspace variation produced by the bolt. The attached photo shows a simulation of how I would do it. Bolt lug to the cartridge recess is the actual dimension to measure.
My V-block in the photo is close to a correct way. To make it better would involve honing a radius on one of the surfaces that would be larger than any on the lug. When the Krags were built it was already known that a radius in the corner would resist cracks from fatigue and also make the cutters last longer. That radius is not very big on Krags but it is there and not to be gauged from, you want to gauge from the flat part of the lug.
Just went back and looked at your photos but did not see the one I saw recently with a star stamped on it. I got a feeling that was one inspected all over and not just critical areas, kind of like a star gauged barrel. I think star gauged barrels were done when fresh tooling was installed and extensive gauging was done. Personal opinion is a star gauged barrel is not necessarily the most accurate. Maybe 10 or 20 barrels later the tooling settles down and actually does a better job.

Back to your on to something there, IF there were some variation in headspace that was addressed during final assembly, I really doubt it was by just going through a bucket of random bolts. IF there was any final fitting is was with known bolts that were plus and minus from a mean dimension and designated by a now unknown mark.

I started a discussion on Krags on another forum, particularly about head space and timing or clocking of the barrel.   (You need to Login
So far a few interesting comments and some guesses. I doubt the Springfield A guys were measuring to 50 millionths of an inch.

---