We examine the operation of an assembly rule to account for the struct
ure of rodent communities in the species-rich southwestern deserts of
North America. The rule specifies the functional group from which the
species should come, rather than the identity of individual species in
the assemblage; it specifies that each species entering a community w
ill tend to be drawn from a different group until each group is repres
ented, and then the rule repeats. Here we tested for operation of the
rule using three data sets: 1) for 115 sites from a small region of Ne
vada we analyzed 11 species of granivorous rodents distributed among t
hree functional groups with different foraging strategies: bipedal het
eromyids, quadrupedal heteromyids, and quadrupedal non-heteromyids; 2)
for the same sites, we added folivores and insectivores to the three
functional groups of granivores to analyze all 14 species of rodents p
resent; and 3) for 202 sites dispersed across the southwestern U.S. we
analyzed 28 species of granivores distributed among three functional
groups: bipedal heteromyids, quadrupedal heteromyids, and cricetids. W
e used 1000 Monte Carlo simulations of random community assembly to pr
oduce frequency distributions for the expectations of the null hypothe
sis to test if observed communities followed the rule overall. Then we
compared observed frequencies of particular combinations of functiona
l groups with those expected by chance. We demonstrate that the rule a
pplies to different functional groupings of desert rodents, and to ass
emblages at different spatial scales. We show that the neutral model u
sed in the simulations is robust and appropriate. We suggest that the
mechanism underlying this rule is interspecific competition, which ope
rates by affecting the probability that species in different functiona
l groups differ sufficiently in resource utilization so as to be able
to coexist. Our results demonstrate the usefulness of an approach that
incorporates both deterministic and stochastic processes of community
assembly.