This paper uses examples from Kenya and the Sudan to argue that the scale at which we pitch our analysis when trying to identify the basic properties of pastoral ecosystems may not be appropriate when it comes to recommending policy measures to secure the continued viability of pastoral herding. Pastoral households are always parts of large-scale economic and social structures. In many cases, such integration has provided the basis for the continued viability of pastoral adaptations. In other cases, the changing nature of local economies and social relations following from integration into large systems, has threatened the viability of pastoral herding or led to increasing differentiation within and between local communities. This paper raises the issue of how we most fruitfully define and delimit ecological and social systems in different local settings. It also questions conventional approaches to community-based natural resource management, which are now attracting widespread international attention.