Environmental and biotic factors regulate species abundance and the co
mposition of ecological communities. However, it is difficult to demon
strate principles of community assembly in nature due to observational
limitations and historical effects. Numerical simulation can be used
to evaluate whether hypothesized mechanisms produce a pattern seen in
natural communities. Most models of many-species communities assume si
milar behavior by all individuals in a population. Here we describe an
agent-based model of many-species communities in which interactions a
mong individual agents are subjected to Darwinian selection (Holland's
Echo model). To evaluate the response of the model to evolutionary me
chanisms, we present a simple version of the model, and describe an ex
periment which evaluated community assembly patterns with and without
selection for genome-mediated interactions, at various levels of invas
ion. Increased invasion rates decreased population size, but increased
species richness and evenness. Interactions acted upon by natural sel
ection resulted in greater population sizes and lower species richness
and evenness than random interactions. Genotype abundances deviated f
rom the expectation of equal abundances, and were affected by both the
invasion rate and nature of inter-agent interactions. Species abundan
ce patterns indicate that communities formed by random and selective i
nteractions follow different assembly rules, and communities formed un
der high invasion rates showed no coherent community-level pattern of
genotype abundances. Thus, both invasion rate and type of interspecifi
c interaction regulate the assembly of species into communities. (C) 1
997 Elsevier Science B.V.