Using computer solutions of the relevant differential equations, many authors have determined the velocity for which traveling domains exist in a semiconductor with a region of negative differential mobility. Recently, Gunn presented a topological argument leading to a closed form equation for the velocity of a finite amplitude domain. For the case of a domain shape controlled by diffusion we show that Gunn's argument is based on a necessary, but not sufficient, condition that a certain singular point be a center. This leads to an incorrect expression for the domain velocity for all but infinitesimal amplitude domains. The topological character of each solution as the velocity parameter is varied is also described.