The importance of competition in structuring herbivore species assemblages is widely surmised, but rarely demonstrated. One view is that past competition over evolutionary time has resulted in niche divergence, such that competition has only a minor influence on extant populations (Owen-Smith 1985), or at least is manifested only intermittently (Owen-Smith 1989). Sinclair (1985) suggested that risk of predation has an overriding effect on species associations. However, Prins and Olff (1998) considered competition to be pervasive within grazing ungulate assemblages, such that species of closely similar size rarely coexist.

Part of the problem is that competition among large herbivores arises largely indirectly via vegetation modification or ‘sward capture’ (Murray and Illius 1996, 2000), rather than through overt interference. Smaller ungulates have the potential to out-compete larger species by depressing vegetation biomass below that needed to meet the greater absolute food requirements of the latter (Illius and Gordon 1987). On the other hand, the vegetation impacts of the bigger species can alter vegetation structure such that habitat conditions are changed for other species. However, the habitat modification need not be detrimental. A reduction in grass height could improve food access and dietary quality for smaller species better adapted to exploit short grasslands, leading to interspecific facilitation rather than competition (Vesey-Fitzgerald 1960; McNaughton 1976; Prins and Olff 1998). Nevertheless, despite the short-term gains in nutritional intake that may result for these small species, consequent increases in population abundance have not been observed (Sinclair and Norton-Griffiths 1983).