3D displacement issues with order=3

I'm trying to extend the linear combination example to 3D. During the debugging, I noticed that I couldn't solve a simple displacement problem in 3D with "order=3" or "order=2" integrals. The gist contains the code I'm trying to run

  https://gist.github.com/wd15/8274429e3b4179c3d74f

The gist only has displacements with no linear combination or periodic boundary conditions. The image produced seems to be badly distorted without the expected symmetry.

Should this work with an "order=3" integral or does it need to be "order=1"? Does this require changing the solver parameters in some way?

Thanks.

-- Daniel Wheeler