Our physical experiments investigate vertical axis rotation of long rigid blocks of continental crust during oblique divergence. Experiments use matchsticks embedded in a 2-cm-thick layer of silicone gel that overlies a rubber sheet. A new device uses two motors to create oblique divergence, resulting in transtensional deformation in the rubber sheet and silicone gel. During experiments, single matchsticks behave like material lines because they rotate into parallelism with the direction of oblique divergence. Matchsticks in arrays, however, rotate away from parallelism with this direction. The sense and rate of rotation of these matchsticks in arrays depend on the geometry of the array, the scaling of experiments (using the Argand number), and the spacing of the matchsticks. Strain incompatibility between the rigid matchsticks and the deformable silicone builds relief on the surface of experiments. Gravity-driven flow of silicone away from zones of positive relief causes instability in the orientation of matchsticks in arrays. These dynamic instabilities are consistent with the predictions by Lamb <1994>. Our results have two tectonic implications. First, long, continental crustal blocks may not rotate into parallelism with the direction of oblique divergence. Second, rigid rotation of crustal blocks may create dynamic topography in the rheological layers of the Earth.