Pastoral man interacts with domestic herbivores (3) by control of populations (species and numbers) in time and their disposition in space. Various management strategies have evolved over the millennia shaped by the characteristic of the physical environment. Nomadism, in the example of the Sahel, was cyclic--seasonal following the rains. In more arid, less predictable environments, nomadism is often completely opportunistic. The objective of nomadism is to harvest the 'rich' (high quantity and quality) resources which, in the Sahel, were also reasonably reliable in time (rainy seasons) and space (the movement of the ITCZ). The selection criteria for interaction (3) is survival and the historical evidence is that these pastoral societies and systems have been remarkably resilient (Coughenour et al. , 1985; Hadley, 1985; Lamprey, 1983; Sinclair and Fryxell, 1985).

   No doubt there was variation in numbers with climate, but the system embodied sufficient resilience to persist, i.e. the system was stable (Honing, 1973; Noy-Meir, 1974b; Walker and Noy-Meir, 1982). Such stability can occur only if there are regulating interactions or negative feedbacks that tend to reduce the forward reaction. Interaction (4) is one of the negative feed- backs operating within grazed systems and ecosystems. Herbivores are harvested to provide meat, milk, and blood. These three items can comprise > 70% of the energy available to pastoral man, with milk usually being the largest component (Coughenour et al., 1985). Because herbivore products provide almost the total resource harvested, and these products are energetically rich, this provision represents a considerable demand on the herbivore population. Even in a system which is maintenance- rather than production- oriented, this is a limiting interaction. There is a limited number of livestock that can be safely harvested given low reproductive rates and the variability of climate. The rates of increase of domestic stock are considerably less than that of the vegetation, and herbivore density cannot closely track the biomass of pasture. Because the grazer and the grazed operate on different scales of time they are not closely coupled. This 'slack' in the system has been reported for other grazing systems (e.g. Caughley et al., 1987) and is associated with stability or persistence.