[wterrell]

I disagree with the notion that the marks were applied before assembly of the barrel to the frame. Such process would be totally impracticable, cost prohibitive, and the refuse rate would be astronomical.
It costs quite a bit of money to produce barrels and frames with the tolerances required to accommodate such an assembly process and would be totally unnecessary whenever other assembly techniques with better, quicker results, which are 'no-brainers' can be conceived and executed by any blacksmith.
Regarding the slight misalignment of the witness marks, steel has memory. If you apply torque to steel and mark its position at the time that it is torqued and constrained, and re-measure a year later, you will find that the steel has begun to retreat. This is part of the process which you are observing approx. 70-100 years after torque.
A practice in the oil field whenever disassembling pipe is to deliver a sharp shock to the joint of assembly and the joint will give to pressure and 'break'. Whenever you have torque pressure and repeatedly apply sharp shock (such as explosion of gunpowder) this will aid in the reverse of torque.
Any production technique as illogical as the assembly, marking, disassembly, and reassembly would have to be documented to be believed. It is too fantastical.