The purpose of this work is to find the region of necessary and sufficient conditions for diffusion instability on the parameter plane (τ, d) of the Gierer–Meinhardt system, where τ is the relaxation parameter, d is the dimensionless diffusion coefficient; to derive analytically the dependence of the critical wave number on the characteristic size of the spatial region; to obtain explicit representations of secondary spatially distributed structures, formed as a result of bifurcation of a spatially homogeneous equilibrium position, in the form of series in degrees of supercriticality.

Methods. To find the region of Turing instability, methods of linear stability analysis are applied. To find secondary solutions (Turing structures), the Lyapunov– Schmidt method is used in the form developed by V. I. Yudovich.

Results. Expressions for the critical diffusion coefficient in terms of the eigenvalues of the Laplace operator for an arbitrary bounded region are obtained. The dependence of the critical diffusion coefficient on the characteristic size of the region is found explicitly in two cases: when the region is an interval and a rectangle. Explicit expressions for the first terms of the expansions of the secondary stationary solutions with respect to the supercriticality parameter are constructed in the one-dimensional case, as well as for a rectangle, when one of the wave numbers is equal to zero. In these cases, sufficient conditions for a soft loss of stability are found, and examples of secondary solutions are given.

Conclusion. A general approach is proposed for finding the region of Turing instability and constructing secondary spatially distributed structures. This approach can be applied to a wide class of mathematical models described by a system of two reaction–diffusion equations.