Coevolution, the process of reciprocal evolutionary change between interacting species, is thought to have played a significant role in shaping almost all life on Earth (Thompson 2005). Some of the most important steps in the diversification of life have probably been made through coevolutionary processes. For example, the fundamental leap from prokaryote to eukaryote cell design is thought to have evolved through symbiotic relationships. The endosymbiotic theory of eukaryote origins, which was largely developed by Lynn Margulis, postulates that cell organelles (particularly mitochondria and chloroplasts) are derived from the engulfment of one prokaryote (bacterial) cell by another (Margulis and Sagan 2002). Chloroplasts are thought to have once been free-living photosynthetic bacteria, and mitochondria were once aerobic heterotrophs, oxygen-dependent organisms incapable of producing their own food. How they gained entry to larger cells is uncertain, but they could have been engulfed as prey or even through evolution of a parasitic relationship. But once inside, the possible benefits to both the engulfer and engulfee are not difficult to envisage: Photosynthetic symbionts could supply nourishment, and as the world became increasingly oxygenated, anaerobic cells would benefit from aerobic symbionts, which used oxygen to their advantage. In turn, residing within a larger cell may provide a safe haven from other would-be predators. Presumably, the relationship between hosts and symbionts eventually became so interdependent that it was impossible to separate one from another, finally resulting in a single organism. Lines of evidence for the endosymbiotic theory are many and include the fact that many enzymes and transport systems in the inner membranes of chloroplasts and mitochondria are most similar to those found in prokaryote cells. Chloroplasts and mitochondria also divide through a binary fission process (not mitosis), which is most like bacteria; similar to prokaryotes, the DNA found within them is also in the form of circular molecules and is not associated with histones or other proteins as it is in the nucleus of eukaryotes. Finally, mitochondrial genes are more similar to those of free-living bacteria than they are to genes in the nucleus of the cell in which they reside.
Symbiotic relationships continued to play significant roles in the diversification of life. For example, colonization of land by early plants, which lacked roots, may not have been possible without the aid of mycorrhizal fungi. Fossil evidence indicates that even the earliest plants were associated with mycorrhizae, which help with the absorption of nutrients. In addition, many chemical defense compounds produced in plants may not be produced by the plants themselves but by the fungi often obligately associated with them (Herre et al. 2005). Tropical seas, which are so poor in nutrients, would probably be comparatively lifeless were it not for the extraordinary relationship between coral polyps and dinoflagellate algae. The polyps provide “housing” and waste products that dinoflagellate algae use in photosynthesis. In return, polyps supplement their daily food intake with photosynthate produced by resident algae. The enormous success of the corals, which are among the largest nonhuman-made structures on Earth, has allowed them to provide most of the structural heterogeneity and nutrients on which many tropical marine species have become so dependent. Even the process of digestion, which most people take for granted in their daily lives, cannot take place without the aid of endosymbiotic bacteria. Although it may be hard to believe that coevolution did not play an important role in some of the most important biological leaps in the explosion of Earth's biodiversity, there have been very few explicit tests of its putative role. Part of the reason for this is that coevolution is notoriously difficult to demonstrate, especially for events that took place millions of years ago or that involved the interactions of numerous species.
What is coevolution?
To show that organisms have coevolved, scientists must demonstrate the process of reciprocal evolutionary change between interacting species, driven by natural selection (Thompson 2005). This definition of coevolution is a modification of Daniel H. Janzen's (1980) definition, which Janzen put forward to halt the injudicious use of the term, which had often been used to describe a host of relationships that had probably not coevolved at all. Prior to Janzen's definition, the word coevolution had been used to explain all manner of symbiotic relationships in which organisms were closely associated but for which there was no direct evidence for reciprocally driven evolutionary change. For example, Janzen questioned the use of the term when it was assumed that the traits of certain mammal-dispersed fruit had coevolved with a particular mammal's dietary requirements. He postulated that a mammal could enter a new habitat with its dietary preferences already established and would begin to feed on the fruits of plants that fulfilled those requirements. Seeds of those plants could evolve adaptations to the gut of the new herbivore, but unless the herbivore evolved adaptations to the new seeds, the relationship would not be coevolutionary.
The English naturalist Charles Darwin (1809–1882) was the first to envisage the reciprocity of coevolution, where changes in one species drive changes in another species, which in turn drive changes in the first species. He wrote about his ideas in On the Origin of Species (1859), where he envisaged “how a flower and a bee might slowly become, either simultaneously or one after the other, modified and adapted in the most perfect manner, by the continued preservation of individuals which presented slight deviations of structure mutually favorable to each other” (pp. 94–95). Darwin was referring to red clover flowers, which are normally visited by bumblebees with long proboscises, whose nectar was mostly inaccessible to honeybees with shorter tongues. He imagined the scenario of bumblebees becoming locally rare and changes in honeybee morphology, such as proboscis length, that would enable them to take advantage of the new resource by better matching the floral morphology of red clovers. But he also envisaged that floral morphology might also change to make honeybee pollination more efficient.
Darwin was a masterful natural historian who had an excellent understanding of how intricately adapted flowers influenced the evolutionary path of their pollinators and vice versa. This was best demonstrated in his first book after Origin of Species, a detailed account of how floral parts are sculpted by their pollinators over evolutionary time. He described how minute floral folds and convolutions are in fact adaptations to Page 195 | Top of Articlepollinator behavior and morphology. In particular he planted the seed of how the extremely long nectaries of flowers evolved and how these could often be the result of coevolution. For an example of this, he used the Madagascan star orchid (Angraecum sesquipedale), an orchid he had never seen in the wild.
This orchid is unusual because it has an absurdly long spur (in excess of 30 centimeters [12 inches]), at the bottom of which a few drops of nectar can be found. Orchid pollen is found in sacks called pollinaria, and these sacks have sticky attachments that adhere to various parts of their pollinator's anatomy when the pollinator contacts the flower's reproductive parts. Darwin postulated that only a large moth with its head pushed hard up against the flower, in its effort to drain the last drop of nectar, could remove the pollinaria. His example suggested that very long-tongued moths would not have to push their heads up against shorter-tubed flowers to get the nectar, and as a result the pollinaria of such flowers would not be removed and the moths would not deposit pollen on their stigmas. Thus, longtongued moths exert a selective pressure on flowers to evolve corolla tubes or spurs that are longer than the tongues of the moths. But at the same time, moths that have short tongues cannot reach all the nectar found at the base of very deep flowers. To access all the floral nectar, selective pressures on moths would favor tongues longer than the tubes of flowers visited. Darwin recognized that “there has been a race in gaining length between the nectary of Angraecum and the proboscis of certain moths” (1877, p. 116), where as one species evolves greater length, it forces the other species to evolve greater length and vice versa. This line of logic led Darwin to make his bold, and at the time frequently ridiculed, prediction that the Madagascan star orchid was pollinated by an enormous hawkmoth with a tongue that matched the orchid spur in length. Unfortunately he never lived to see his prediction tested because the hawkmoth was only found forty years later, long after his death.
Scientists have conducted several tests of the prediction that selection favors increased tube length in flowers pollinated by long-tongued insects, but perhaps the most famous is the experiment by L. Anders Nilsson (1988). Nilsson artificially shortened the spurs of some plants in a natural orchid population and found that plants with shortened spurs had fewer pollinaria removed and less pollen deposited on their stigmas. Although no conclusive tests have yet shown that long-tubed plants drive selection of longtongued pollinators, coevolutionary arms races have been used to describe many other directional trends, such as the observation that brains of predators and their prey get larger over time, or that their bodies become increasingly adapted to increased speed over time. Even mollusk shells thicken over evolutionary time in response to evolutionary advances made by their predators. In a similar way to the nuclear arms races of the cold war, these so-called evolutionary arms races can produce an escalation in traits, but the fitness of one organism in relation to the other remains more or less the same.
The relationship between plants and insects is often not mutualistic. Every plant on the planet is probably consumed by insects, but many plants evolve mechanical (e.g., spines) or often chemical defenses to protect themselves from herbivores. To eat plants with chemical defenses, insects must evolve resistance to the plants' natural insecticides. In one of the most important works on coevolution, Paul R. Ehrlich and Peter H. Raven (1964) examined dietary patterns of butterfly larvae. They found that the insects' diets were restricted to a particular subset of often-unrelated plants, which had similar chemical defense mechanisms. They interpreted this pattern as evidence for the manner in which plant defenses and insect resistance coevolve. When an insect evolves a new defense against a particular plant chemical, all those plants possessing that chemical suddenly become available as a new food resource. This can open a host of new niches for the insect to exploit, allowing the insect to diversify and perhaps even speciate. Plants, at the same time, are continually evolving new chemicals to counter the evolution of insect resistance, which may have also resulted in diversification and speciation in plants.
The close associations of angiosperms with their pollinators and herbivores have led some biologists to postulate that diversification through coevolutionary relationships between plants and animals may have been responsible for the enormous species diversity in both of these groups. If these two groups played major roles in each other's speciation rates, Page 196 | Top of Articlethey would have rapidly diversified at the same time. The fossil record suggests that major diversification started in the angiosperms during the Cretaceous. Conrad C. Labandeira and John J. Sepkoski Jr. (1993) calculated the proliferation of insect families from about 25 million years ago to the recent past and found that the number of insect families increased logarithmically with time, with no apparent acceleration in the Cretaceous. Although this can be taken as evidence against coevolutionarily induced radiations of these taxa, the methods used by Labandeira and Sepkoski have been severely criticized. Rapid radiations are often characterized by diversification at higher taxonomic levels, such as within genera. Thus looking for a diversification at higher taxonomic levels, such as insect families, may not be appropriate if insect diversity had radiated at lower taxonomic levels. Other tests of this hypothesis have examined whether insect-pollinated plant taxa are more diverse than plant taxa pollinated by abiotic means (e.g., wind). Michael E. Dodd, Jonathan Silvertown, and Mark W. Chase (1999) compared related branches of the angiosperm phylogeny, where one branch was insect pollinated but the other was abiotically pollinated. In contrast to that of Labandeira and Sepkoski, the results for Dodd and colleagues showed that insect-pollinated branches of the angiosperm phylogeny are more diverse than abiotically pollinated branches, suggesting that associations with insects did increase angiosperm diversity. This is not necessarily coevolution, however, because, in return, scientists also need to test whether plant diversity played a role in the diversification of insect taxa. So the jury is still out on whether reciprocal associations between insects and plants were responsible for each other's diversity.
Although scientists are still unsure whether coevolution was responsible for reciprocal insect and plant radiations, many believe that coevolution is implicated in the evolution of sex itself. And sexual reproduction has been shown to speed up rates of evolution and adaptation, which may in turn have consequences for speciation rates and ultimately the diversity of life on Earth. Evolution and maintenance of sex is one of the thorniest biological debates in science, but one of the main contending hypotheses is the red queen hypothesis, which is based on host—parasite coevolution. Here, parasites evolve virulence toward the most common host genotype in a population. In the next generation, however, a previously rare genotype will be the most resistant to parasite virulence, and hosts with this genotype should increase in frequency until it becomes beneficial for parasites to evolve virulence toward them instead. Thus, the environment for both hosts and parasites is constantly changing every generation, so that parasites and hosts, like Lewis Carroll's Red Queen, may Page 197 | Top of Articlecontinually run a cyclical arms race instead of a directional one. Because sex allows recombination of genotypes, it can recreate genotypes that were lost in the past because they were so disadvantageous. These recreated genotypes may be needed in future bouts of coevolution to combat parasite virulence or alternatively to combat host resistance. The same cannot be said for asexually reproducing organisms, because when their disadvantageous genotypes are lost they may take hundreds of generations to rebuild through mutational processes. Strong evidence for the red queen hypothesis comes from Mark F. Dybdahl and Curtis M. Lively (1998), who used snails and their trematode parasites to show that parasites were adapting to the most common host genotype in a population and that, as a result, hosts underwent cycles of their genotype frequencies.
Hosts and their parasites provide some of the best examples of coevolution (not only red queen coevolution), such as the classic work on the myxoma virus and their rabbit hosts, which were introduced without their parasites to Australia where they rapidly became pests. After the virus was introduced, the rabbit population was nearly eliminated but then recovered. By isolating both rabbit and virus strains before the virus was introduced (pre-introduction rabbits and pre-introduction viruses), scientists showed that rabbit recovery stemmed from both a reduction of virus virulence and an increase in host resistance. This was demonstrated after viruses taken from Australian rabbits, several years after the virus introduction, failed to kill as many pre-introduction rabbits as the pre-introduction viruses killed. Similarly the evolution of rabbit resistance was demonstrated by “evolved” rabbits having a lower mortality rate than pre-introduction rabbits after both were exposed to pre-introduction virus strains.
The geographic mosaic of coevolution
Despite the examples of coevolution given above, most coevolution probably does not involve just a single species pair. In any population, several species may exert selective forces upon one another. Over time, the composition and abundance of communities may fluctuate, and at the same time, the strength of selection imposed by each species will fluctuate as well. Similarly, a single species may interact with Page 198 | Top of Articleseveral species across its geographic range so that the outcomes of interactions can vary geographically. Since the 1980s, these ideas have been carefully formulated in a series of books and papers, mostly written by John N. Thompson and colleagues (see Thompson 1994 and 2005, and the references therein). One of Thompson's model study systems is yucca plants and their close relatives, which are visited by moths that frequently pollinate as well as parasitize them. In a landmark paper, Thompson and Bradley M. Cunningham (2002) showed that in certain parts of the plant's range the relationship between plant and moth was mutualistic. In other parts of the range, however, the relationship was antagonistic, and in still others the relationship was commensalistic. The divergent outcomes of this relationship appear to be the result of species composition, as well as the presence, absence, and abundance of copollinators at each site.
Yet another example of how geography and community context can play a decisive role in structuring coevolutionary outcomes can be seen in the three-way relationship between crossbills, lodgepole pines, and red squirrels. In the central and northern Rocky Mountains, red squirrels (Tamiasciurus hudsonicus) are the main seed predators on lodgepole pines (Pinus contorta), and consequently these pines show many defensive character traits adapted primarily to squirrel predation (Benkman, Holimon, and Smith 2001). Because crossbills (Loxia spp.) exert comparatively little selective pressure on the pines here, pines show no adaptations to specifically prevent predation by crossbills. Crossbills have nevertheless evolved bill adaptations specifically to pry apart the scales of pinecones. In contrast, some areas peripheral to the Rocky Mountains have no red squirrels, and so the major seed predators in these systems are crossbills (Benkman et al. 2003). In these systems, the plants have adapted to crossbill predation by evolving thicker scales where most of the seeds are located. In turn, the crossbills here have evolved, through coevolutionary processes, even larger and more decurved bills to pry apart the thickened scales. Thus, geographically variable selective pressures have caused both matches and mismatches of phenotypically complementary traits in interacting organisms.
Using examples such as these, Thompson has constructed what he calls the geographic mosaic theory of coevolution. Here, the outcomes of interspecific interactions vary across the geographic landscape, in some places forming coevolutionary hot spots and in other places coevolutionary cold spots. Mediated by geneflow, trait remixing, and community context, this selection mosaic of coevolutionary hot spots and cold spots will pulse and shift across the landscape. These divergent and morphous patterns of selection should have enormous consequences for how biological communities evolve and function and ultimately on how Earth's biodiversity is organized. Only since the formulation of the geographic mosaic of coevolution hypothesis have scientists really begun to appreciate the pervasiveness of coevolution on everyday life and the ramifications it has on applied sciences such as biological control, agriculture, human epidemiology, and conservation.
Darwin, Charles. 1859. On the Origin of Species by Means of Natural Selection; or, The Preservation of Favoured Races in the Struggle for Life. London: John Murray.
Darwin, Charles. 1877. On the Various Contrivances by Which British and Foreign Orchids Are Fertilised by Insects. 2nd edition. London: John Murray.
Herre, Edward Allen, Sunshine A. Van Bael, Zuleyka Maynard, et al. 2005. “Tropical Plants as Chimera: Some Implications of Foliar Endophytic Fungi for the Study of Host Plant Defense, Physiology, and Genetics.” In Biotic Interactions in the Tropics: Their Role in the Maintenance of Species Diversity, ed. David F. R. P. Burslem, Michelle A. Pinard, and Sue E. Hartley. Cambridge, UK: Cambridge University Press.
Margulis, Lynn, and Dorion Sagan. 2002. Acquiring Genomes: A Theory of the Origins of Species. New York: Basic Books.
Thompson, John N. 1994. The Coevolutionary Process. Chicago: University of Chicago Press.
Thompson, John N. 2005. The Geographic Mosaic of Coevolution. Chicago: University of Chicago Press.
Benkman, Craig W., William C. Holimon, and Julie W. Smith. 2001. “The Influence of a Competitor on the Geographic Mosaic of Coevolution between Crossbills and Lodgepole Pine.” Evolution 55(2): 282–294.
Benkman, Craig W., Thomas L. Parchman, Amanda Favis, and Adam M. Siepielski. 2003. “Reciprocal Selection Causes a Page 199 | Top of ArticleCoevolutionary Arms Race between Crossbills and Lodgepole Pine.” American Naturalist 162(2): 182–194.
Dodd, Michael E., Jonathan Silvertown, and Mark W. Chase. 1999. “Phylogenetic Analysis of Trait Evolution and Species Diversity Variation among Angiosperm Families.” Evolution 53(3): 732–744.
Dybdahl, Mark F., and Curtis M. Lively. 1998. “Host—Parasite Coevolution: Evidence for Rare Advantage and Time-Lagged Selection in a Natural Population.” Evolution 52(4): 1057–1066.
Ehrlich, Paul R., and Peter H. Raven. 1964. “Butterflies and Plants: A Study in Coevolution.” Evolution 18(4): 586–608.
Janzen, Daniel H. 1980. “When Is It Coevolution?” Evolution 34 (3): 611–612.
Labandeira, Conrad C., and John J. Sepkoski Jr. 1993. “Insect Diversity in the Fossil Record.” Science 261(5119): 310–315.
Nilsson, L. Anders. 1988. “The Evolution of Flowers with Deep Corolla Tubes.” Nature 334(6178): 147–149.
Thompson, John N., and Bradley M. Cunningham. 2002. “Geographic Structure and Dynamics of Coevolutionary Selection.” Nature 417(6890): 735–738.