We investigated the spatial pattern of A. raddiana in the Negev desert of Israel in order to gain insights into the factors and processes driving the dynamics of this species. Using a scale-dependent measure, the ring statistic, we analysed both patterns observed in the field and time series of spatial tree distributions produced by a simulation model. In the field, random spacing was the predominant pattern observed. However seedlings were clumped on small scales. We ran the model under two contrasting scenarios representing hypotheses that explain the clumping of seedlings and the random distribution of trees. One hypothesis is that there is spatial heterogeneity in seed distribution, germination and seedling mortality, but that these heterogeneities are not correlated with each other in space. The second hypothesis assumes a correlation between these heterogeneities leading to areas suitable for establishment. However, the suitability of the sites is temporally variable. Furthermore, the second hypothesis assumes density-dependent tree mortality due to competition. Both hypotheses lead to spatial distributions that are in qualitative agreement with the patterns observed in the field. Therefore, the classical view that a clumped seedling distribution and a random pattern of older trees is due to clumped regeneration and density-dependent mortality may not hold for Acacia trees in the Negev.