In considering the interaction of ecosystems and environment, May (1974) wrote: 'it is to be emphasized that although patterns may underlie the rich and varied tapestry of the natural world, there is no single, simple pattern. Theories must be pluralistic. ' The ecological theory available to interpret the pattern and process of arid ecosystems is substantial. The contributions of, particularly, Noy-Meir(1973, 1974a, 1980,1981,1985), Shmida et al. (1986), Westoby(1980) and Crawford and Gosz(1982) provide a unifying framework that is supported by the voluminous and diverse observations reported in various synthesis volumes (e.g.Goodall and Perry, 1979,1981;Evenari et al. , 1985,1986).

   The most persuasive theory is that of Noy-Meir(1980), who has proposed an 'autecological' hypothesis rather than the 'ecosystem' hypothesis to interpret observations of arid landscapes. The latter implies a system in which there is a high level of connectedness, strong interactions, regulatory or catalytic feedbacks between the components (populations) and the abiotic environment. The organization or structure of such a system, its persistence and functional properties, are distinctive and differ from those of the individual component populations. This model has guided ecologists working with communities that enjoy a high level of constancy or predictability of abiotic environment (e.g. Odum,1983,1985; Pimm,1982;Odum and Biever,1984; Menge and Sutherland, 1987).

   The 'autecological hypothesis' recognizes that, in arid ecosystems, the population dynamics of plants and animals is determined largely by the physical environment and little affected by interactions such as competition and predation. In particular arid ecosystems are water-controlled in an environment where the input of water is highly variable in time, space, and amount. Consequently ecosystem activity is similarly highly variable in time and space. Therefore, in contrast to the 'ecosystem' model, there is a low level of connectedness, weak or intermittent interactions between populations, and few if any regulatory feedbacks.