INTRODUCTION Ant communities are usually referred to as being highly interactive and exhibiting clear intra- and interspecific relationships (Oliveira & Della Lucia 1992). Although several papers point to interspecific competition as an important structuring factor in ant communities (Levings & Franks 1982; Fellers 1987; Savolainen & Vepsäläinen 1988; Andersen & Patel 1994), there are few studies testing this role of competition. The relative role of biological interactions in the determin- ation of species diversity thus remains largely unknown in ant communities. An adequate method for evaluating the relative roles of local processes, such as competition, on species diversity regulation is the analysis of the relationship between local and regional species richness (Terborgh & Faaborg 1980; Cornell 1985a,b; Ricklefs 1987; Hugueny & Paugy 1995; Griffiths 1997). Two different patterns may be detected in this analysis: (i) unsatu- rated, a linear relationship between local and regional species richness; and (ii) saturated, an asymptotic relationship between these two variables, which means that the local species richness increases proportionally with regional species richness up to a level when local species richness becomes independent from regional species richness (Cornell & Lawton 1992; Cornell 1993; Cornell & Karlson 1997). The absence of a relationship between local and regional species richness is also interpreted as evidence of saturation (Caley & Schluter 1997; Srivastava 1999). Although some authors prefer to use the term saturation only when the pattern is due to species interactions (Cornell 1993), we use the term sensu Terborgh and Faaborg (1980) to describe the pattern of the general relationship between local and regional species richness, dis- regarding the processes involved in generating the pattern. Saturated patterns may be generated by strong biotic interactions, such as interspecific competition (Cornell 1985a; Cornell & Lawton 1992). When the competi- tion is intense and widespread, it may act by limiting the species number present in a given community. In this case, there is a maximum species number able to coexist and, when the community reaches such a number, the local species richness remains constant, even if the regional species richness increases. Although species saturation is frequently associated with interspecific competition, non-interactive processes, such as species pool exhaustion and stoch- astic equilibrium (Cornell 1985b; Cornell & Lawton 1992; Srivastava 1999) may produce the same pattern. Species pool exhaustion corresponds to the restric- tion of the number of species adapted to a given local environment and it is determined by the degree of Austral Ecology (2001) 26, 187–192 Processes involved in species saturation of ground-dwelling ant communities (Hymenoptera, Formicidae) SANDRA M. SOARES, 1 JOSÉ H. SCHOEREDER 2* & OG DESOUZA 1 1 Departamento de Biologia Animal, Universidade Federal de Viçosa and 2 Departamento de Biologia Geral, Universidade Federal de Viçosa, Viçosa, MG, CEP 36571-000, Brazil (Email: jschoere@mail.ufv.br) Abstract The species saturation hypothesis in ground-dwelling ant communities was tested, the relationship between local and regional species richness was studied and the possible processes involved in this relationship were evaluated in the present paper. To describe the relationship between local and regional species richness, the ground-dwelling ant fauna of 10 forest remnants was sampled, using 10 1 m 2 quadrats in each remnant. The ants were extracted from the litter by using Winkler sacs. Using regression analyses, an asymptotic pattern between local and regional species richness was detected. This saturated pattern may be related to three processes: (i) high interspecific competition; (ii) habitat species specialization; or (iii) stochastic equilibrium. It is concluded that non-interactive processes, such as stochastic equilibrium and habitat specialization may act as factors regu- lating species richness in this community. The predominance of locally restricted species, in all sampled remnants, seems to indicate the occurrence of a high degree of habitat specialization by the ant species. This result is evi- dence for the hypothesis that community saturation has been generated by non-interactive processes. Although ants are frequently described as highly interactive, it is possible that interspecific competition is not important in the structuring of ground-dwelling ant communities. Key words: habitat specialization, interactions, local diversity, regional diversity, species richness, stochastic equilibrium. *Corresponding author. Accepted for publication September 2000.