I got rid of most of the rocking and I think I know how to get rid of it all but I have not yet got around to it.
I was actually looking for why my table did not crossfeed smoothly so I took the table and casting below the table off. Then I coated the dovetail on the casting with Prussian Blue and manually rubbed the casting against the tabletop dovetail, on the side opposite the gib. I found the contact was bad, i.e. spotty and only happening for about 3/4 of the dovetail length. I took an Arkansas stone that was left over from when I sharpened wood carving chisels and stoned the casting surface. I'd spend about 20 minutes on it then reapply blue and check it again. After most of a full day I had it stoned so about 1/2 of the dovetails made contact. It actually looked like there was two planes machined into the dovetail and I was getting contact on all of one of the planes. (The line of demarcation was too straight to have resulted from my stoning it.) When complete the table moved much better.
Then while I had it all apart I decided to use the bluing to check the casting- to-lathe-way dovetail. It was just about as bad. Basically the dovetail surface on the casting (the side opposite the gib) seemed to be bowed. Depending upon how I held the casting against the ways I could get either the tailstock or headstock end to make sliding contact but not both. I spent the next day with the stone and made the sliding contact much smoother and longer, but not perfect. When I put it all back together again I noticed that the table left/right motion was smoother and that there was less table rocking. That makes sense because as the table-to-lathe-ways dovetail gets straighter there is less opportunity for the table to rock on the high spot in the center of the dovetail.
I have a surface grinder that is not running. As I get time I work on it and it will be running by year end. I now plan to buy a cup wheel for it and then too fully grind both dovetails. I expect that may fully resolve the rocking and stiffen the machine enough to eliminate some of the mill chatter.
While taking things apart I also noticed that there are two socket head cap screws that hold the under-table casting tight to the plates that support the crossfeed lead screw thrust bearings. Those two cap screws were not recessed into the slots designed for them and instead were on washers that bridged the slots. The washers were collapsed into the slots and that allowed the table more motion. I fixed both things at the same time so I don't know how much looseness each contributed.