Introduction

The role of spatial population structure in promoting cooperation and mutualism has recently received much interest in a number of theoretical ecological and evolutionary studies (e.g., Nowak and May 1992; Hammerstein and Hoekstra 1995; Killingback and Doebeli 1996). It has also been emphasized in the context of the origin of life for essentially the same reason: it is a means to establish coexistence of potentially competing template replicators, thus allowing increased capacity of information storage and transmission by the population as a whole (see Maynard Smith and Szathmáry 1995 for review). There are three known, detailed model approaches:

• Structured deme-type models (Wilson 1980; Michod 1983; Szathmáry 1992)

• Replication–diffusion systems as modeled by cellular automata (Boerlijst and Hogeweg 1991b)

• Group selection of replicators encapsulated in compartments (Szathmáry 1986; Szathmáry and Demeter 1987; Maynard Smith and Szathmáry 1993)

The motivation for these studies originates with a seminal paper by Eigen (1971) arguing (1) that primitive genomes must have been segmented (consisting of physically unlinked genes); (2) that these unlinked genes must have had the tendency to compete with one another; and (3) that, consequently, some mechanism ensuring their coexistence was needed. Eigen saw the hypercycle (Figure 7.1), a system of cyclically interacting molecular mutualists, as fulfilling this role. However, in a spatially homogeneous setting, the hypercycle is vulnerable to parasitism: a cheating replicator that does not give catalytic aid to any member of the cycle can kill the cycle off, provided it receives more catalytic help from the cycle than the member that it competes with in the first place.