Official Luthiers Forum!

What you might do, in the experimental spirit, is bolt some channel iron (on edge of course) on the TOP of the table at the first t-slot in from each edge. Make the channel iron of the same cross section that could then be used under the table following Bob's "W" suggestion. Better yet, do this virtually with your FEA.
FEA (FiniteElementAnalysis) is a wonderful tool that I miss greatly since retiring from my day job. We had ANSYS integrated with ProE which allowed quick changes in the solid model to run "what ifs". ANSYS has very powerful and quick meshing capability but then it's not cheap by any means.

@Nelson - the sag on this machine is .030-.035" over 48". It isn't static, it is worse with the gantry away than home but I can tell that the endplates deform more in that position. I've had this nagging suspicion that perhaps the aluminum steel rod rails have settled into the deformed position and wonder if a loosen and retorque with the table center jacked would help. The 4 12mm hardened steel rods should be helping. If they are helping, then the table is worse than it looks with them I should be able to simulate the benefit of supports in the FEA but Solidworks Premium is pretty limited and it is time consuming to figure out all the things that cause runtime to go infinite or meshing to fail. I definately want to try bracing it externally before major surgery just to see how far that goes.

@BobG - The W thing is cool and a good idea (THANKS !). There is currently 2mm clearance between the gantry bottom and the base table at both ends of travel. Avoiding that due to the new bracing would shorten travel on each end nearly 2". I have headstock setups on the away end, with the shaft hanging out over the end of the table, so that is just something else I'll need to solve for that scenario.

Rand, something else to try when attaching stiffeners to the top or bottom side of the table would be to preload the stiffeners. Just a flatwasher spacer or two at each end of the stiffener on top of the table would apply some preload and, conversely, a spacer at the middle if the stiffener is bolted to the underside. A better approach would be to use washers of varying thickness at each bolt location in the stiffener to provide a "smoother" preload, if that makes sense. Pieces of graduated shimstock on each side of a bolt location would give a similar effect.