So, I've stripped the vice down – that is the jaws off and the fixed block removed. A bit of grit under that as well. The edges of the underneath were also a bit rough so I took the edge off.
I spent some time measuring the fixed block with and without the jaw on.
The fixed block is quite rough on the side facing the jaw. The depth is various by around 0.03mm across the surface with a distinct channel in the vertical direction around the middle. of the block. There is more variance at the bottom of the block than further up.
Despite this and the earlier reported irregularity of the jaw, with the jaw on there is a drop of about 0.01mm across the top of the jaw and block assembly and about 0.02 across the bottom.
I checked out the height of the sliding surfaces on the surface table and they seemed spot on. This measurement was more dubious as the vice is longer and nearly as wide as my surface table!
After cleaning and re-assembly, I put the vice back on the mill and clocked the fixed jaw. This now shows a deviation of 0.025mm (one thou) from left to right.
So my conclusion is that the problem was mainly down to dirt in the vice. one thou out across 5 inches is more than enough for just about everything I do. If I have an unexpected need for better than that I wlll take the fixings off and clock it in the normal way.
Many thanks for all the advice.
Iain