As I mentioned a couple of days ago, it is a 1/8" thick sheet of steel, 27.5" x 27.5". I drilled 1/2" holes 1/2" from the corners, and mounted the 3" total locking casters that I use for ScopeRoller there. The locking casters are necessary, because when you start using the digital setting circles, you do not want any motion relative to the surface; it otherwise would make the sensors on the ground board inaccurate.
I would not want to roll this over gravel or grass, but on pavement and concrete it works just fine.
Friction alone is not enough to keep everything on that sheet of steel. I took a piece of 1" square aluminum tube that I had lying around, cut it into three pieces, and drilled through the top surface so that I could recess the screws holding the aluminum to the steel sheet. You can see one of the edges of the ground board here.
There is a bit of flex in the steel sheet, but it isn't very obvious, and I am not particularly worried about the problem getting worse, so I think I will leave it well enough alone for the moment. I mentioned on the Second Light posting that I had loosened and tightened one of the truss tube connectors, and I thought that explained the collimation problem. Indeed, when I recollimated, I found that I was significantly off because of that. I look forward to rolling Baby out in her carriage tonight.
UPDATE: Some more pictures for Dennis Steele at DobSTUFF, who built this telescope.
A bit more detail of the square tubing used to hold the ground board in place: