Summary 1. Demographic processes, particularly the recruitment of seedlings, are critical to the long-term survival and re-establishment of plant populations. In many perennial grasslands successful recruitment is a rare event that requires the simultaneous favourable alignment of several environmental variables. 2. Although a few empirical studies have examined demographic processes in perennial grasslands, most of these studies investigate the impact of only one or two environmental factors in isolation. Those studies would lead us to expect that seed availability, irrigation and burning, enhance recruitment rates whereas the presence of established competitors substantially lowers recruitment rates. But how these factors, or recruitment filters, act in concert and how they influence the different components of the recruitment process such as seedling emergence, survival and growth is not understood. 3. For this reason we explore the relative and combined influence of four factors – seed availability, irrigation, competition by the established vegetation, and fire – on the recruitment process of a perennial grass (Stipagrostis uniplumis ) in semi-arid savanna. We used a full factorial experiment to determine how these four factors influence emergence, growth, survival and flowering of seedlings. 4. Our results show that all of the factors affected some stage of the recruitment process but not all significant effects contributed to enhancing the final number of recruits. 5. We found enhanced seed availability to significantly enhance rates of seedling emergence but to play no further role in the recruitment process. Competition by the established vegetation exerted a strong negative effect on every step of the recruitment process. Irrigation had a minor effect on recruitment. It enhanced rates of flowering and survival during the growing season but the effect of the stimulus was short-lived and did not influence later stages of the recruitment process. 6. In contradiction to existing theory, seedling recruitment does therefore not seem to be limited to above average rainfall years. Fire acted similarly to the removal of competitors; that is, it positively affected seedling emergence, growth, flowering and survival. 7. Synthesis . We conclude that recruitment is not limited to above average rainfall years but triggered by events that open up the grass canopy and reduce the abundance of competitors. Specifically it appears that periodic fires enable the recruitment of new individuals into the population. This study shows that fire can play an important positive role in individual turnover of semi-arid perennial grass populations. Key-words: demography, emergence, full factorial experiment, intraspecific competition, irrigation, recruitment limitation, seed addition, seedling survival, semi-arid savanna, Stipagrostis uniplumis |