It was suggested that a simple application of the composite model to a system of differently spaced walls predicted that an abundance of dislocations would be deposited at the interfaces of walls which faced broad channels, as compared with those which faced narrow ones. By tentatively assuming that this surface dislocation density could be somehow related to dislocation fluxes into the walls, it was concluded that the walls should move in a direction which was opposite to that previously assumed.

On Moving Dipolar Dislocation Walls. R.Sedlacek: Physica Status Solidi B, 1997, 201[2], R5-6