1 Spatial pattern analysis in Namaqualand desert plant communities: evidence for general positive interactions N. S. Eccles;, K. J. Esler; & R. M. Cowling Abstract This paper investigates of the spatial arrangement of individual plants in mapped plots in two desert communities in the winter rainfall region of South Africa. In both communities there was a very strong tendency towards clumped patterns when all plants were considered together. There was also a predominance of clumped patterns when the most abundant species in both communities were considered individually. When the arrangement of the most abundant species was considered relative to the arrangement of all other individuals at the within-clump scale, there was a high frequency of positive associations in both communities (62% and 65%). We speculate that these patterns represent a combination of seed dispersal strategies that favour clumped patterns and a predominance of positive interactions between plants in both of the communities. When specific pairwise associations between the most abundant species were considered at the within-clump scale, differences were apparent between the two communities. In the short strandveld community neutral associations predominated, while in the medium strandveld, neutral and positive association accounted for equal proportions of the associations. This between-plot difference was also apparent when the volumes of plants were related to an index of neighbourhood competitiveness. In the short strandveld there were no significant relationships while in the medium strandveld there were some weak (but significant) relationships. These differences were not altogether unexpected. If we assume that plants in the medium strandveld are generally longer-lived, then interactions between plants are likely to develop over a longer time and, therefore, are likely to be stronger. Keywords: Coexistence, Diversity, Mutualism, Winter rainfall deserts Introduction The winter-rainfall Succulent Karoo of southern Africa is an unusual desert ecosystem. Local and regional species richness levels are higher than any other desert system with as many as four times the number of species as equivalent sized areas of North American winter rainfall deserts (Cowling et al. 1998). It is estimated that there are some 4489 species in an area of 112 000 km 2 , with about 40% of these species being endemic to the biome (Hilton-Taylor 1996). It is also unusual in terms of the growth forms that dominate the system. Short- to medium-lived succulents (and in particular leaf succulents) predominate and the geophyte flora is unusually rich. In contrast, there is a comparative lack of grasses and large shrubs (Milton et al. 1997). Despite the unusual structural composition of these communities, the ratio of functional diversity to species diversity is relatively low, suggesting a great deal of niche overlap and species redundancy (Cowl- ing et al. 1994). This poses the question: how is this diversity, with its apparent redundancy, maintained in the system? The answer to this is particularly elusive in the context of communities occurring on the coastal sand plains where small- scale heterogeneity is very limited and as such, the potential for diverse ‘habitat niches’ (Hubbell & Foster 1990) is also limited.