Species evolve in ways that improve (or at least maintain) their viability. Communities, it seems, do not have the same process: there is no evolution of a community or ecosystem as a unit. But species may evolve in ways that are primarily influenced by ecology: by their relationship with one or a few other species, and in turn these others evolve also. This is coevolution, a term that was originally coined by Ehrlich and Raven after observing plant-herbivore interactions.

There are variations within coevolution: one species may converge evolutionarily on another that changes little (for example, in some forms of mimicry). Two species may diverge until they no longer interact (character displacement, maybe driven by niche partitioning). Species may respond to such a broad range of other species that "coevolution" is diffused into a general reaction to the biotic environment, or "evolution": examples here include the response of herbivores to the whole ecosystem represented in the word "savanna," or the intricate ways of life that have evolved within "reef" ecosystems. Nevertheless, the concept of coevolution is very important, and in a broad sense may underlie some major adaptive radiations.

Models of rapid speciation depends heavily on the chance events that split off isolated populations. These events, by their very nature, tend to be aspects of the physical environment, such as changes in climate or topography. But organisms interact with other organisms, which form part of their biotic environment. Would our view of the evolutionary process change if we considered biotic aspects too?

There are two broad categories of coevolution: antagonistic and mutualistic. (Enlarging on this theme: for each there must be a balance or an asymmetry such that one species is evolutionarily the "prey" and the other is the "pursuer", except in the case of runaway selection for mutual advantage. It would be to the advantage of the "prey" or "target" species to escape by quantum speciation, except that I don't see at the moment how the species could accomplish that. In most cases of very strongly linked biotic coevolution, one species is likely to track another very closely. Thus in coevolution, there is probably a built-in bias toward gradual evolution, in both or all species involved. Now it depends how coevolutionarily linked any given community is: there's no need to suppose that the links are simple. Perhaps in a biotically interlinked community with strong internal biotic relationships, the dominant style of evolution could be coevolutionary and gradual. Perhaps evolution could be more punctuated in loosely linked communities. Maybe it depends whether biotic interactions or physical changes are dominant for particular groups of organisms at particular times and in particular ecosystems.)

If a major selective force on a species is its relationship with another species, then one can envisage an process in which they evolve toward a mutual evolutionary stability. In fact, if the relationship is adversary, such as a predator-prey relationship, or a parasite-host relationship, achieving mutual stability is necessary for their continued coexistence. The mutual stability might take the form of character divergence so that there is no longer a coevolutionary relationship.


Predator-prey relationships

Asymmetry. There is an asymmetrical pay-off in those predator-prey relationships where the prey is killed. The predator pays for failure by short-term lack of food: it gets a second chance, usually. The prey pays for failure with its life. This asymmetry, with corresponding direct feedback into fitness, might possibly produce more visible and extreme adaptations in prey species than in predators. Bakker, in an attempt to show that coevolution is imperfect, documents an apparent failure by early Cenozoic predators to "keep up" with the advanced locomotory systems of early Cenozoic herbivores.

Arms races. There is bound to be dynamic coevolution between any predators and prey that coexist for a while. One would then predict "arms races" as a natural consequence, especially in those cases where the prey is not killed (e.g. in many plant-herbivore interactions). Thus defenses may take the form of toxins, and herbivores will endeavor to "crack" the code of the toxins to make them into metabolites. If successful they may then use the toxins (monarch butterflies, opisthobranch gastropods).



The size of herbivores: other things being equal, creatures move faster at large size, and become predator-proof at larger size. (Is this the underpinning of Cope's Law? Because if the suggestion is true, predator size should then track prey size too, and most things are either predators, or prey, or both.)

What is the role of predation in evolution? Stanley supposed that predation acted to promote diversity, misusing Paine's work on keystone species: actually, predation merely maintains diversity. Vermeij made a much better case for predation encouraging diversity.

Lost races? Trilobites vs. fishes and/or cephalopods?


- often at one trophic level



: is subtly antagonistic. The result is that mimics never come to outnumber their models. But by the same process, a model that is once mimicked may never be able to escape except by its own extinction. Although code-breaking takes place, we're looking at an asymmetrical evolutionary race again.




Mutualism must have its origins in antagonism.

Return to Evolution topics

Return to Geology Department home page