Update: I disassembled the table and put it on parallels on the surface table – to my surprise the error seen on the mill had disappeared and the table surface was now flat to within +/- 0.02mm.
I also noted that the x-gib was polished to a shine at the very lower edge on the sliding face indicating it had probably never been in proper contact resulting in the distortion to the table surface.
I was unable to obtain a replacement gib from the supplier so I had a go at re-scraping it. I superglued a 2 thou shim to the non-sliding face and embarked on a process of blueing the face of the gib, re-assembly and then sliding the table to-and-fro by hand (I left the leadscrew off to allow feel).
The first round of bluing confirmed that the gib was indeed only in contact with the lower edge of the dovetail, after many passes & a few hours of blueing & careful scraping I eventually ended up with a series of high spots across the whole contact face. To finish, I scraped in some deeper 'scuffs' to help with oil retention and reduce 'stiction'.
Now the table has been re-assembled, the action is very smooth with no rocking and the error previously seen across the table width has now been reduced to about +0.02mm and this occurs at various points across the table rather than being a pronounced hump in the middle.
Thanks for all the advice.
Dave