Turing-type reaction-diffusion mechanisms have been suggested as methods of pre-pattern formation in many biological situations. These models have been subjected to criticism on the ground that in the case of zero flux boundary conditions the solution patterns are extremely sensitive to changes in the initial conditions and in the domain size and shape.
This paper considers the situation in which the domain is permitted to interact with its environment (itself a biologically plausible assumption), which requires the use of boundary conditions of non-zero flux. Through linear stability analysis and computer simulations the authors show that the patterns that emerge are much less sensitive to changes in the initial conditions and domain shape and size than those arising in the zero flux case.