Roubo leg joint - constraining the top?

[Prashun Patel]

I am designing a table that will use a Roubo leg joint (double tenon through the top).
If I add rails connecting the front and back legs, then isn't the top effectively constrained? Won't the rails eventually crack? The only thing I can figure is that locating them sufficiently low enough allows the legs to pivot and accommodate the expansion. If this is true, then does this pose a requirement on how the legs are connected to the rails?

Any help appreciated.

(disclaimer: This image is taken from Wilbur Pan's blog, Giant Cypress)

[Sam Stephens]

I would be concerned as well. Locating the apron at the top may be too restrictive; towards the bottom of the legs may allow the legs to distort to accommodate the movement of the top, but not sure if that's desirable depending on the width of the top/anticipated movement. Can the apron float, at least at the back leg?

[Art Mann]

If the bench you are building is very similar to the one in the picture, I don't think you will have any problems because the difference in the coefficient of expansion of the rails and the top is so little that the pivot angle you mentioned will be next to zero.