Assembly rules and competition in desert rodents.

North American desert rodents have long been a model system for the study of community structure and dynamics. Field experiments and nonmanipulative observations have provided consistent evidence that interspecific competition influences spatial patterns and dynamic processes of community organization. Some of the evidence comes from assembly rules, which document how the species that coexist in local communities differ in body size, trophic morphology, functional group, or taxonomy more than would be expected had they been assembled by some random process (Kelt and Brown 1999). One assembly rule proposed by Fox (1987; see also Fox and Brown 1993; Fox 1999; Kelt et al. 1999) is based on functional groups. It hypothesizes that species are assembled in such a way that each functional group tends to be filled before an additional species is added to an existing functional group. Communities that conform to the rule are termed “favored states,” whereas other combinations are termed “unfavored states.” So if the rule is operating, we should observe more favored states than expected by chance. Although Fox’s assembly rule has been applied to a variety of systems (1999), its application to desert rodent communities has generated controversy. Fox and Brown (1993) showed that there were more favored states than expected by chance in two large data sets: a sample of 115