Ehrlich and Raven in their 1964 study coined the term “coevolution” and since then it has grown into a topic of great interest for both evolutionary biologists and ecologists alike. Coevolution refers to the adaptation of one population or species to another that leads to a reciprocal adaptation in the first species. When coevolution occurs between a single pair of species it is considered “tight” coevolution. In contrast, diffuse coevolution occurs when selective pressure is generated by multiple species interactions (Janzen 1980). Whether coevolution is tight or diffuse is dependent upon the ecology of the organisms involved. The nature of ecological communities, however, illustrates the interdependence of species on each other. Darwin himself proposed the idea of the tangled bank, where species interactions are intricately interwoven and connected. These ideas prompt the question of how coevolution proceeds when local populations interact with a broad range of species, many of which are exerting some selection pressure on an interaction (Thompson 2005).
One approach to answering this question lies in the study of ecological networks. An ecological network is one in which species (nodes) are connected to one another through specific interactions (resource-consumer, mutualistic, host-parasite, etc.). Traditionally the most commonly studied ecological networks have been food webs (e.g. Lindeman 1942, May 1972, Pimm 1984, and for more recent studies see Rooney et al. 2006, Petchey et al. 2008). Recently interest in studying other ecological networks, such as mutualistic networks, has increased (Bascompte 2009, Bascompte 2010). As the study of ecological networks has become more established in the literature it has become clear that this approach can be useful in studying evolutionary as well as ecological processes (Bascompte 2009, Loeuille and Loreau 2005, Loeuille 2010).
A network approach allows for the study of how coevolutionary dynamics act within a framework of multiple species interactions and facilitates a shift in thinking from tight to diffuse coevolutionary theory. Pair-wise interactions have traditionally been the basis for ecological and evolutionary studies, but the new emphasis on communities has demonstrated the important role of networks in multispecies interactions (Proulx et al. 2005). Studies of coevolution in the pair-wise manner may suggest that, through the application of the evolutionary “arms race” analogy, reciprocal selection and coevolution should favor extreme specialization within the pair. When coevolution is applied in the context of ecological networks we see drastically different effects of reciprocal selection. In mutualistic networks, for instance, complementarity and convergence leads to the so-called “coevolutionary vortex” (Thompson 2006) affecting the growth of the network and its implicit structural properties. Networks are generally comprised of repeated patterns of small groups called motifs. One common motif in food webs is the tri-trophic food chain (Bascompte 2009). Finding these motifs within the network of interest should allow evolutionary ecologists to easily study important coevolutionary modules within the larger network scheme.
Two major misconceptions about coevolution are that (1) it leads toward highly specific one-on-one interactions and (2) that in species rich communities coevolution generates diffuse assemblages that cannot be generalized (Bascompte 2009). A network approach to coevolution helps to dispel these misconceptions by showing the lack of evidence for reciprocal selection within one-on-one interactions (Joppa et al. 2009). The study of various ecological networks and non-ecological networks has shown as well that many properties of networks are universal (Jordano et al. 2003, Proulx et al. 2005).
In John Thompson’s (2009) review “The coevolving web of life” he argued in favor of the use of a network-based approach in conjunction with the geographic mosaic theory to study coevolution. He also asks five “big” questions about ecological patterns that are shaped by coevolution among species including one about the structure of ecological networks. Of those questions, “How does the structure of reciprocal selection change during the assembly of large webs of interacting species?” is particularly relevant to the study of ecological networks. In other words how does coevolution among species affect the shape of the network, and in turn how does the shape of the network influence coevolutionary interactions? Coevolution between multiple species is most likely to shape the structure of the webs through the addition, and subsequent assimilation of species into the community. Adding new species (new nodes) into the network will generate new links among species altering the patterns of links within the whole network and possibly leading to the loss of other nodes (species extinction). In turn the structure of the network determines the manner and strength with which its component species interact.