with particular reference to herbivory
COEVOLUTION WITH PARTICULAR REFERENCE TO HERBIVORY
Of all the extant organisms in the world, it is believed that terrestrial plants and their natural enemies constitute more than forty percent. Moreover, plants exhibit a remarkable diversity of supposedly defensive characteristics including trichomes, spines, silica, secondary chemical compounds, temporal avoidance of enemies, and structures along with chemicals that attract predators of their natural enemies. In addition, the exploitation of the plants and their defences is facilitated by a vast number of behavioural, morphological and physiological adaptations by herbivores
Accounting for this diversity has been a major area of research for nearly a century. The seminal article, attributing this diversity to coevolution, was published in 1964 by Ehrlich and Raven. They suggested plants and herbivorous insects evolved reciprocally by the following events: Plants, through occasional mutations and recombinations, produced a series of chemical compounds not directly related to their basic metabolic pathways. Some of these compounds, by chance, serve to reduce or destroy the palatability of the plant in which they are produced. Such a plant, protected from the attack of phytophagous animals, would in a sense have entered a new adaptive zone. Evolutionary radiation of plants might follow.
If a new recombinant or mutant appeared in a population of insects that enabled individuals to feed on some previously protected plant, selection could carry the line into a new adaptive zone. Here it would be free to diversify in the absence of competing herbivores. Ehrlich and Raven (1964) emphasised the importance of the reciprocal selective responses between ecologically linked organisms.
Since 1964, studies have questioned Ehrlich and Ravens postulates. Due to the nature of evolutionary study, ideas are only as strong as the background in the literature; that is, acceptance by the scientific community depends upon its knowledge. In time people learn more and previously weak theories become more feasible. Alternatively, and more so in science, accepted work in time becomes disregarded (example; until the 1950s geologists believed in static continents, now all believe in plate techtonics and continental drift). The significance of this is that any theory published is only speculation of what is happening in these interactions. The knowledge is blind in that historical findings leading to these assumptions are not concrete. What happened in the past might be a different picture to what we have envisaged so far.
Thompson (1999) has proposed that there are crucial components to coevolution. These need to be recognised before we can fully understand coevolution. Firstly, phylogenetic studies are providing five kinds of data important in interpreting the historical context of coevolving interactions. 1) Shared traits. Phylogenetic studies are allowing us to evaluate which traits of interacting species were already present in the hosts ancestors. This allows us to determine whether traits are coevolved or merely a trait exhibited as a consequence of the organisms genotype. For example, Yucca plants provide a source of food for host specific Yucca moths, with which they are believed to have coevolved. Examining the phylogenetic trees of these moths elucidated this. Most moths in this family (Prodoxidae) exhibited host specificity (Davis et al 1992). Before this technology, people would have assumed the specificity of the Yucca moth to be a product of the coevolution.
This brings up a useful comment by Vermeij (1994). Almost all inferences about coevolution are derived from the existence of trends in the expression of traits that function during interactions between species. Evolutionary trends have often been found by analysing ancestor-dependant relationships within monophyletic groups, or clades. Although many trends are best sought this way, others cannot in principle be detected within single clades and instead arise when ecologically and functionally comparable clades replace each other through time.
2) Unique traits. The Yucca moths as described above have tentacles on their mouthparts used to hold pollen for later active transfer to floral stigmas. Using phylogenetic studies, it has been found that the ancestors to these moths did not have tentacles, suggesting a coevolutionary adaptation.
3) Relative malleability of traits. Regardless of selection intensities, some traits may be more malleable than others. Recognising this enables us to discriminate between organisms that appear to be evolutionarily constrained (low malleability) and those that appear to be evolutionarily dynamic (high malleability). Those that have low malleability of a particular trait are not necessarily fixed and using phylogenetic studies can determine how fixed they are.
4) Multiple origins of interactions. Selection for convergence of traits, especially in mutualisms, can confound interpretation of the coevolutionary process in the absence of a good phylogenetic template.
5) Relationships between traits and patterns of diversification. Our understanding of how evolution of new traits have shaped diversification in interacting taxa has been enhanced by phylogenetic studies. There are currently four hypotheses on the interrelationship between species interactions and diversifications of taxa.
 Parallel cladogenesis (phylogenetic tracking) suggests host specific parasites speciate with their hosts simultaneously but do not cause speciation. They may or may not be coevolving through coadaptation but the host speciation is independent of any coadaptation. That is, coevolution is not a cause of reciprocal speciation.
 Sequential evolution. This theory suggests parasites track their hosts speciation, but do so in a much more general way. The theory assumes they are not coevolving. Hosts undergo periodical diversification due to other factors, then the parasites colonise the new host.
 Escape and radiate coevolution. This is the formal version of Ehrlich and Ravens (1964) hypothesis as explained earlier. This theory does not predict strict parallel cladogenesis.
 Diversifying coevolution. This group of hypotheses suggest; 1) populations of a species evolve to specialise (often on different species) as a result of reciprocal local adaptation. 2) Hybrids among the specialists populations are at a selective disadvantage, thereby favouring reproductive isolating mechanisms.
Coevolution demands some degree of reciprocal specialisation among interacting species if the interaction is to affect the fitnesses of individuals and favour evolutionary reciprocal change. Evolutionary arguments are often based upon this fact.
Opposition to coevolution note that coevolution occurs when only a few species interact. This rarely happens, most interactions are widespread, so coevolution is uncommon or, at best, diffuse. This view is questionable, coevolution of a plant with its pollinators does not stop coevolution with its herbivores or dispersal agents. Nevertheless, the plant-pollinator coevolution may well be checked before it reaches optimum by the other interactions of the plant.
Plants undergo herbivory. Generally, they are a suitable source of nutrition for animals. Interpreting Van Valens (1973) Red Queen hypothesis in this context means that any plant will become extinct unless it constantly evolves adaptations to the local attack upon it. By that rationale, it must develop suitable defence to the herbivory to survive. Along a continuum,
No defence Max. Defence
starting at no defence and traversing to complete defence, depending on the specific situation, plants in the middle sections will be selected for. Plants with little defence mechanisms (A) will be over-consumed, therefore becoming extinct. Plants which resort to utilising the maximum available resources (C) for defence will survive but not be selected for. Plants towards the middle of the continuum (B) will have spent more resources and energy than C in reproducing thus increasing its relative fitness. The most successful plant will be the one that has stumbled (note: the plant makes no conscious decision of position) upon the most appropriate position in the continuum. Moreover, as Van Valens work predicts, that plant will not be successful for long if it ceases its evolution, the parameters about it will change and select for a new phenotype.
So the plant has to provide enough protection to survive as a species, (not necessarily itself- altruism) but maintain as much resource as possible committed to its primary evolutionary role, reproduction. A dynamic trade-off has been put in place.
The cost of this protection, or resistance, varies in the methods employed by the plant. Rausher (1996) has identified three types of cost to the plant, each different cost constraining the plants evolution in some way;
1. Allocation cost. Plant resources redirected to resistance characteristics are withheld from other characteristics that enhance survival, growth and reproduction. These costs are revealed if the predator if taken from a habitat, the genotypes conferring high resistance have a low relative fitness. For example, May Berenbaum grew numerous lines of wild parsnip (Pastinaca sativa) in a field and a pest free glasshouse. The parsnips had varying levels of genetically induced furanocoumarins (feeding inhibitors) Outdoors, the parsnips with higher levels of furanocoumarins had a higher yield and fitness. In the glasshouse (with no predators), the best yield was from parsnips with less furanocoumarins, the higher concentration parsnips had an inappropriate amount. This 2° pool of resources can be as much as 10% or more of a plants resource budget (Rosenthal and Berenbaum 1992)
2. Ecological cost. This arises when genotypes employing a resistance mechanism against one predator becomes susceptible to another because of it. For example Crucifers have 2° compounds (glucosinolates) present which act to deter feeders, which they do. The cabbage white butterfly (Pieris brassicae) uses this as a chemoattractant and revels in the plants attempt to dissuade predators.
3. Recently discovered, the trade-off between resistance and tolerance is the third cost. While tolerance can be viewed as a type of defence against natural enemies, it differs from resistance in that resistance is the average amount of damage experienced by a genotype for a given abundance of natural enemies. Tolerance is the average amount by which the fitness of a genotype is reduced for a given amount of damage. Fineblum and Rausher (1995) showed that genotypes that were resistant to herbivores that damage the apical meristem of I. purpurea were less tolerant of damage by those herbivores and vice-versa. This type of cost varies from allocation cost in that allocation costs are manifested in the absence of natural enemies, while the cost of resistance in the form of tolerance is not.
A distinction between pairwise and diffuse coevolution has recently been proposed (Iwao and Rausher 1997). Pairwise coevolution occurs between a plant and two enemies if the trajectory of coevolution between the plant and one of its enemies does not depend on presence of the other enemy, otherwise the coevolution is diffuse.
Another debate currently going on is the question of antagonistic versus mutualistic coevolution. Antagonistic coevolution is the classical theory as proposed by Erhlich and Raven (1964). This idea was the benchmark, the theory of mutualistic coevolution was developed afterwards suggesting a more harmonious arms race between the interacting species. Mutualistic coevolution was posed, as a mechanism after compensatory growth was observed in plants after herbivory. Vail (1992) showed, in limited cases, that plant responses to herbivory in terms of compensatory growth meant a mutualistic relationship was going on. The term mutualism implies the plants compensatory growth ensures a higher yield than non-herbivorised plants. Of course, the results would not be the same if the plant were not eaten in evolutionary time. I question the mutualistic hypothesis, perhaps we are just viewing a frame from antagonistic coevolution. Alternatively, the plant has teleologically learned that it will probably be eaten upon and makes use of this. For instance, it may use the herbivore as a seed dispersal agent. Both descriptions appear to be linked to pairwise coevolution, which suggests (because now diffuse coevolution is being mathematically modelled) that the theories or the definitions will be altered or combined.
Finally, the whole process is becoming less cloudy. New findings are occurring more often. Of particular help to the study is the worldwide genome characterisation on many organisms currently going on. This technology will help us to identify more gene for gene interactions such as found in the Hessian fly / Wheat interactions. (Schoonhoven et al 1998)
To progress in the study of coevolution, the scientific world needs to treat it as a true eco-evolutionary science. The first step of showing selection acting upon certain traits must be built on. We must begin to fully understand when, where and why the process of coevolution shapes communities.
Word Count: 2000
Rory Auld BS7RA
 Davis, DR. Pellmyr, O. Thompson, JN. (1992) Biology and systematics of Greya Busck and Tetragma, new genus (Lepidoptera: Prodoxidae)- smithson. Contrib. Zool. 524 1-88
 Ehrlich, PR. Raven, PH. (1964) Butterflies and plants: a study in coevolution. Evolution. 18 586-608
 Fineblum, WL. Rausher, MD (1995) title. Nature 377 517-520
 Iwao, K. Rausher, MD. (1997) Evolution of plant resistance to multiple herbivores: Quantifying diffuse coevolution. Am. Nat. 149 (2) 317-331
 Rausher, MD. (1996) Genetic analysis of coevolution between plants and their natural enemies. Trends in Genetics. 12 (6) 212-217
 Rosenthal, GA. Berenbaum, MA.(eds.) (1992) Herbivores: Their interactons with secondary plant metabolites. Edition II. Academic Press, San Diego, U.S.
 Schoonhoven, LM. Jermy, T. van Loon, JJA.(1998) Insect-plant biology. From physiology to evolution. Chapman and Hall, London
 Thompson, JN. (1999) The raw material for coevolution. Oikos. 84 5-16
 Vail, SG. (1992) Selection for overcompensatory plants responses to herbivory: A mechanism for the evolution of plant herbivore mutualism. Am. Nat.139(1) 1-8
 Van Valen note: get Van Valens 1973 red queen hypothesis reference.
 Vermeij, GJ. (1994) The evolutionary interaction among species: Selection, Escalation and Coevolution. Annu. Rev. Ecol. Syst. 25 219-236