As organisms evolve to take better advantage of their surroundings, they may come into competition. A predator may compete with its prey, or two species eating the same plant may compete with one another to find it. With only a limited amount of resources to go around, living things have to adapt not only to climate, geography, and other slow-changing variables but also to the more rapid evolutionary changes undertaken by their competitors. Although change in climate can be seen as a comparatively slow change, one's living competitors are constantly evolving, and this requires that both parties keep on their respective toes: Fighting for resources can be a never-ending battle, and evolution ensures that the playing field is rarely level. However, evolution does not necessary breed outright competition, as it is also possible for two species to enter a mutually beneficial relationship whereby they help each other. No species lives in a vacuum, and potential interactions with one's neighbors can be beneficial, costly, or both. The chart below details the possible outcome of species interactions.
Note that predation does not differentiate lions eating antelopes from antelopes eating grass—both are predators eating prey, and prey does not have to possess a backbone or sturdy legs. In each case, both predator and prey can respond evolutionary to one another, and those changes may be physiological or behavioral. The antelopes can evolve to run faster. The lions might respond by evolving to run faster too. The grass, although it cannot run away, can evolve to taste bad to discourage antelopes from eating it. This might force the antelopes to find another type of grass, or their taste buds might change. When two (or more) species begin to respond evolutionarily to one another, this is called "coevolution."
Nature offers countless examples of coevolution. Many of these result from predator-prey interactions. Prey may evolve camouflage defenses, which require that predators become progressively better at detecting them. Walking sticks, for example, are insects that closely resemble twigs and are therefore frequently overlooked by predators. Because walking sticks are difficult to detect, their predators must have more acute vision and a greater ability to discern them from the inedible wood they mimic.
Predators, too, may employ camouflage to position themselves closer to prey. Yellow crab spiders reside in yellow flowers, patiently waiting for tasty insects to land on the flower looking for nectar. Such prey finds a nasty surprise instead. Scorpion fish, an extremely toxic species of fish found in temperate and tropical oceans, closely resemble the rock-covered sea floor. When inattentive prey drifts too close to the lurking scorpion fish, it swiftly gets eaten.
As one species in such a relationship grows more and more camouflaged, the other must get better at detection. More complex camouflage structures result for one, and a more refined nervous system and perhaps keener vision, smell, or hearing develop for the other. The driving forces for these adaptations are the other species, not temperature, geography, or any other variables.
Experiments suggest that camouflage is an effective tactic prey can use to confound predators. Theodore D. Sargent and the team of Alexandra T. Pietrewicz and Alan C. Kamil showed in 1981 that blue jays had a difficult time detecting underwing moths (Catocala) resting on certain backgrounds. The moths have wing patterns that blend in with tree bark, provided that they are oriented on the bark properly. Slight variations in the moths' wing patterns are called polymorphisms . If consistently shown just one polymorphism, blue jays became quite effective at detecting it even when a moth was on tree bark and oriented correctly. But if the blue jays saw different polymorphisms in random order, they never learned. Thus, in this evolutionary arms race, the blue jays have evolved the ability to detect subtle camouflage, but only one type at a time. Adapting to this ability by introducing a polymorphism enables the moths to defeat the blue jays' learning and detection abilities.
The ability of predators to detect hidden, or cryptic, prey can be found elsewhere. For example, John L. Gittleman and P. H. Harvey showed in 1980 that chicks were able to learn to discern camouflaged prey in a comparatively small number of trials. Predators may get better at identifying prey through a variety of mechanisms. They may get better at detecting camouflaged prey against a specific background or they may restrict their searches to more limited areas. In response, prey become more and more devious in their disguises, and each species can exhibit enormous evolutionary pressure on the other as they interact more frequently and grow increasingly adept at duping the other.
Relationships among Species
How do such relationships start? Imagine a worm that can partially conceal itself from its predator, a blue jay. Given that a blue jay has to look for any worm it eats, extra time spent looking for that hidden worm may well be wasted. In a limitless universe of worms, with half being concealed and the other half being unconcealed, the blue jays will never bother to identify the concealed ones, because it will be far easier to just pluck up the easily visible ones. As such, the concealed worms gain a selective advantage over the obvious ones.
But suppose the obvious worms all get eaten up. Then the blue jays have to start eating concealed worms, which they cannot find nearly as fast. So blue jays with better sensing abilities will tend to eat more, grow healthier, and reproduce more. Before long, with an abundance of blue jays with keen vision, the partially concealed worms will start to become scarce, unless they can better conceal themselves.
With a small advantage, that of partial concealment on the part of the worms, an evolutionary arms race has been created. The worms get better and better at hiding, and the blue jays get better and better at finding them. In a competitive world, this arms race can be seen everywhere. For example, trees grow taller to get more sunlight than their neighbors. Their neighbors, in response, grow taller too. For all this growing, however, the amount of sunlight reaching the trees does not change: They are expending increasing amounts of energy competing for the same, unchanging supply of a resource.
Competition of such intensity can be costly. To describe this phenomenon, Leigh van Valen coined the term "Red Queen Principle" in 1973. This term comes from an observation made by the Red Queen in Lewis Carroll's Alice in Wonderland. Alice and the Queen had been running furiously but could not go anywhere. The Queen told Alice, "Here, you see, it takes all the running you can do to keep in the same place. If you want to go somewhere else, you must run at least twice as fast as that!"
The Red Queen principle suggests that competing species may have to allocate more and more resources into fighting one another for a modest or negligible increase in benefit. As each side grows leaner and faster and better able to fight the other, the balance between competitors can be maintained. However, what is to be said of uneven matches? The struggle between bacteria or viruses and mammals is seriously lopsided. Bacteria, as parasites, can infect mammals and live off of them. A long-lived mammal, such as a human, may take twenty years or longer to go from birth to reproducing age, whereas the infecting bacteria may be able to reproduce within a matter of hours. With a faster generation time and many more mutations when they reproduce, bacteria can adapt to different environments and evolve much more quickly. Given this discrepancy, one might be tempted to think that competition between organisms with disparate generation times might always go to the ones that can evolve more quickly. In scenarios like that of lions and antelopes, it seems like an even match. But when bacteria tussle with humans, one might initially think the bacteria should always win.
The Evolutionary Struggle
Fortunately for humans, this is not always so. While the evolutionary arms race gives rise to new structures with which one fights the enemy, it can also give rise to structures that get around the problem of slower generation times. An internal simulation of evolution is an incredibly intricate structure, and it helps illustrate the heights of complexity that an evolutionary arms race can produce. The mammalian immune system has devised a number of strategies that closely resemble a tightly controlled simulation of evolution: The mechanism that generates antibodies (and T-cell receptors) recombines genes far more quickly than does the conventional method of mammalian reproduction. These genes, designed to recognize fast-changing bacterial and viral invaders, can change as fast as their competition.
Like all evolution, bacterial mutations must be beneficial for the bacteria to survive. The genes encoding antibodies do not particularly affect the survival of antibody-generating cells (B cells). But for the system to be effective, the body wants only the cells possessing the genes that can catch up with the bacteria. Thus, after creating an isolated scheme to accelerate mutation of specific genes, the body must create selective pressures to guarantee that it gets only the ones it wants. It does this, too, in the lymph nodes. By keeping little pieces of the bacteria around, the body can select only the B cells that best recognize the invader and discard the rest. Thus, even with lengthy generation times and low mutational rates for the rest of their genomes , mammals are able to simulate the conditions of rapid turnover and high mutational rates inside their own bodies to combat invaders with the same characteristics. Stronger and faster muscles are different manifestations of the arms in question, as is the antibody system. One set makes the organism go faster, and the other makes it selectively evolve faster.
Selectively speeding up evolution is not necessarily restricted to organisms that need to catch up to their competitors. The butterfly genus Heliconius boasts brightly colored wings and produces foul-tasting chemicals to discourage predators. Once a predator eats a Heliconius butterfly, it quickly learns to avoid butterflies with similarly idiosyncratic markings. But what if a butterfly that is not Heliconius can mimic the colors on Heliconius wings? This mimic can enjoy the reputation of being poor prey without actually having to manufacture the foul-tasting components itself.
Of course, this mimicry does not help Heliconius. If a predator comes upon a mimic and finds it tasty, it becomes more likely that a Heliconius might be eaten later, foul taste or not. This is another form of an evolutionary arms race: The mimic gets an advantage at Heliconius' loss. Such mimicry, when a nontoxic species tries to look like a toxic one, is called Batesian mimicry. It is fairly common in the insect world: For example, the hornet moth (Sesia apiformis), the wasp beetle (Clytus arietis), and the hoverfly (Syrphus ribesii) all have the same characteristic stripes as the common wasp (Vespula vulgaris), but only the common wasp has that painful stinger. Each one of the stingerless species usurps an ornery reputation from the common wasp, and as predators learn to eat them, the common wasp suffers more predation.
There are a few potential responses to these Batesian mimics. The predator could get better at discerning mimics from the real thing. As the predator gets better, the mimics too will get closer and closer to the real thing to fool the predator. Alternately, the original, toxic species could change its markings, forcing the mimics to change as well. If the mimics and the toxic species have roughly the same mutation rates and generation times, these changes might proceed at the same rate.
Heliconius has come up with a different strategy. It has evolved a select toolbox of wing patterns from which to choose during development. This pattern tactic resembles the rapid evolution enjoyed by the mammalian immune system— Heliconius can quickly change its wing pattern over a few short generations rather than taking a long time. Its mimics, on the other hand, must slowly evolve to get that precious wing pattern and thus avoid their predators without having to taste toxic.
Not all coevolution needs to be adversarial. Flowers are a coevolutionary adaptation for pollination. Many plants have flowers that are tailored to the needs of a specific insect or bird or bat. Flowers are designed to catch the eyes of certain animals, and the nectar inside is meant to appeal to particular tastes. Bees cannot see red but readily pick up blue, green, yellow, and ultraviolet. Butterflies have decent vision but a poor sense of smell, so they tend to pollinate brightly colored but odorless flowers.
Other structures can keep the relationship between a plant and its pollinator very close. For example, certain flowers have parts designed specifically for the length of a moth's tongue, and only that certain moth species can drink the nectar inside. Common snapdragons are designed so that the flower opens when an object the exact weight of a bumblebee lands on it.
These adaptations are mutually beneficial. The animal provides the plant with pollination, which means the plant can reproduce. The plant provides the animal with nectar, which can feed and sustain the animal. To ensure reliable pollination, the plant evolves to become more and more recognizable by its pollinator, and, conversely, the pollinator gets a steady food source. Each side grows stronger and can reproduce more under the partnership. In such cases, coevolution is an enormous asset to the species involved.
Coevolution of two organisms, therefore, aptly demonstrates that evolution does not have to respond exclusively to nonliving forces. Slight pressure from a competitor, or an ally, can redirect a species. Some pressures, such as predatory ones, can cause the species to invest more in camouflage, detection, or muscle and speed. Others, such as ones that lead to mutualism, may promote structures that will help a species more closely work with its partner, like a plant with its pollinator. Also, different strategies can lead to the bending of the rules, such as getting around the problem of evolution rates. As organisms come into close contact with another, a coevolutionary strategy is the most efficient path to success.
see also Behavior; Behavioral Ecology.
Davies, N. B., and M. Brooke. "Coevolution of the Cuckoo and Its Hosts." Scientific American 264 (1991):92-98.
Gittleman, J. L., and P. H. Harvey. "Why Are Distasteful Prey Not Cryptic?" Nature 286 (1980):149-150.
Sargent, T. D. "Antipredator Adaptations of Underwing Moths." In A. C. Kamil and T. D. Sargent eds., Foraging Behavior: Ecological, Ethological, and Psychological Approaches, ed. A. C. Kamil and T. D. Sargent. New York: Garland Press, 1981.
van Valen, Leigh M. "Patch Selection Benefactors and a Revitalisation of Ecology." Evolutionary Theory 4 (1980):231-233.
When two kinds of organisms exert natural selection on each other so they influence each other's evolution, they are undergoing coevolution. Any two organisms may exert selective pressure on each other. Herbivores exert selection on plants favoring the evolution of defenses, and plant defenses exert selection on herbivores to overcome them. Competitors exert selection on each other favoring superior competitive ability. Pollinating insects exert selection on flowering plants to provide attractants and rewards, and plants exert selection on pollinators for superior pollination service. This reciprocal natural selection is the core concept in coevolution. It may produce ongoing evolutionary "warfare," in which the participants constantly change their weapons or tools, or it may produce a relationship that benefits both participants. When the outcome is beneficial to both, it is called mutualism.
In 1964 entomologist Paul Ehrlich and botanist Peter Raven suggested that these reciprocal changes in physical, chemical, or behavioral traits could be great enough to generate new species. Theoretically, as selection favors changes in each partner, the altered partner could differ from its ancestor enough to become isolated as a new species. For example, if a plant gains protection from its parents' enemies (disease or insects) by producing novel defenses, and if this protection is lost by sharing genes with the parental plant types, then selection should eventually eliminate mating between these two types, resulting in two species where before there was one. Natural selection may then favor enemies capable of colonizing the new plant species, with subsequent reproductive isolation and the formation of additional enemy species. New enemy and plant species are thus formed. Ehrlich and Raven claimed that coevolution may be the major kind of interaction generating the diversity of species on land. While many scientists are skeptical of that statement, the evidence of coevolution is all around us, and many fascinating relationships in nature have arisen from it.
Evidence of Coevolution
Most plants and animals experience natural selection from many sources at once. So it seems unlikely that one organism would be the sole or even the primary selective influence on another. Nonetheless, there are good examples of tightly coevolved relationships (the two participants have a highly specialized interaction). In these cases, the selective advantages gained by responding to one source of selection (the other participant) must outweigh many other factors.
For example, butterflies in the cabbage butterfly family (Pieridae) feed primarily on plants in the cabbage family (Brassicaceae). Members of the cabbage family (cabbage, broccoli, mustards) all share a common set of chemical defenses, called glucosinolates, that are found in very few other plant families. Species in the cabbage butterfly family are capable of feeding on these toxins without harm. According to the coevolutionary view, a mutation long ago in a cabbage ancestor provided that plant with the ability to make glucosinolates, which allowed it to escape the pests plaguing its glucosinolate-free ancestors. But soon natural selection favored butterflies with mutations allowing them to feed on glucosinolates, and these butterflies were able to eat the new plants. Additional mutations in the plants produced new glucosinolates, protecting those plants but selecting for butterflies that could overcome the new chemistry, and so on. In the coevolution scenario, the ability to produce glucosinolates and stepwise responses to evolving enemies resulted in the cabbage family as well as the cabbage butterfly family. If we were to draw cladograms, or evolutionary trees, for co-evolved insects and their host plants, they would be near-mirror images, since each chemical change and speciation event among the plants should have produced one in the insects, and each change in the insects should have produced one in the plants.
Factors Inhibiting Coevolution
Closely matched trees are said to be concordant, an indication of co-evolution between two sets of organisms. Scientists have thus far found few concordant trees involving plants and insects, for at least four reasons. First, it is very difficult to construct such trees, especially for insects, because the fossil record (and even current knowledge about insect diets) is so incomplete. Diets are not preserved in the fossil record. Second, insect and plant evolution are influenced by many things. Most plants are attacked by many different kinds of enemies, and a single defense is unlikely to work equally well against all. And insect success is dependent not only on food, but on weather, escape from predators and disease, and other factors. So plants may not be the single greatest influence on insects or vice versa.
Third, these selective factors interact. The susceptibility of insects to predators, parasites, and disease is also influenced by plant defenses, sometimes in a direction opposite to the way chemistry influences growth and reproduction. For example, gypsy moths grow larger and produce many more eggs when feeding on aspen leaves than on oak. But they are killed readily by a viral disease when they feed on aspen and are protected by oak leaves. So there are conflicting selective forces acting on the insects. The net result is that gypsy moth caterpillars do not distinguish between oaks and aspens consistently. Similarly, plant defenses against their own diseases sometimes inhibit production of defenses against herbivores. This would make coevolution between plants and herbivores very unlikely.
Fourth, herbivores usually do not exert enough selection to favor major changes in the plants. They rarely consume more than a small fraction of their plant food and seldom kill plants outright. Compared with other factors, like obtaining water, nutrients, and light, herbivores are seldom the strongest evolutionary influence on plants. Similarly, competitors infrequently exert the kind of influence on plant neighbors that would produce coevolutionary patterns. Plants exhibit adaptations to competition, including growth responses to the green light reflected from neighbors, and perhaps the production of chemicals toxic to competitors (allelopathy). But there are few, if any, clear cases of mutual adaptation among plant competitors.
But diseases do kill plants frequently and so exert strong selection on plants favoring defense responses specific to the attacking microbe. Scientists have documented many gene-for-gene interactions between plants and pathogens in which a single gene difference between two plants can determine susceptibility to a given microbe. A single gene difference between two microbes can determine which can successfully attack a given plant. One can clearly see evidence of coevolution in these cases, where plants have responded to pressure from microbes with successive genetic and biochemical modifications and the microbes have responded in kind to those changes. A few similar examples do exist for plants and insects, in plant species (e.g., conifer trees, parsnips) with defenses strongly influenced by genes (and less by environment) and insects that can kill them.
Mutually Beneficial Coevolution
Perhaps the most striking examples of coevolution involve mutualisms, in which the participants have exerted selection that makes their relationship increasingly beneficial to each of them. In mutualisms natural selection has favored traits in each participant that strengthen or improve the relationship and its benefits. These interactions contrast with those described above, in which each organism participates at the other's expense.
For example, insects and other animals that transfer pollen among flowers (pollen vectors ) provide a crucial service to the plant while receiving a reward, usually nectar and pollen itself. Because it is disadvantageous for a plant's pollen to be deposited on the flower of another species, natural selection has favored the evolution of traits to reduce these "errors," usually by narrowing the range of species attracted and moving pollen. For example, flowers may produce nectar guides, patterns that reflect ultraviolet wavelengths, making them visible only to certain insects. Others may provide a long, tubular entrance accessible only to night-flying moths with long tongues. The length of a flower's corolla tube is often matched to a particular moth having the same tongue length. This ensures that the pollen will be carried only to other flowers with the same tube length, presumably of the same species. Some flowers provide necessary resources for specific insects, such as oils needed to cement a bee's nest or for mating purposes. In each case, there presumably has been a series of evolutionary changes in the flower (such as corolla tube length) that exerted selection-favoring changes in the pollen vector (tongue length, for example), fine tuning the interaction to mutual benefit.
The evolution of mutualism provides opportunities for deception. For example, many species of orchids produce colorful flowers and odors but provide no reward. They depend on mistakes made by inexperienced bees to get their pollen onto a vector. To ensure that a mistake pays off, the orchid is constructed so that any visiting bee necessarily carries away the pollen in sticky packets called pollinia deposited on its body. The flower is constructed so that when the bee makes a second mistake the pollinium is removed and deposited on the stigma of the second flower. Pollen transfer has to be efficient; terrestrial orchids in temperate North America may only be visited once in a decade.
Some tropical orchids improve their chances of being visited by producing volatile chemicals that are collected by certain bees and used as mating signals. Some orchids may produce an odor that mimics a bee's mating signal, attracting bees that are then disappointed in finding no mate, but carry away a pollinium. In more elaborate coevolved interactions the orchid flowers actually look like a female bee or wasp, with which males attempt to mate. In yet others the flower resembles a male bee, and territorial males attack it. In these latter situations, the pollinia are deposited on the bee when it contacts the flower to mate or fight. All of these deceptive floral adaptations produce a very dependable relationship between the plant and insect (pollinator constancy), but at the insects' expense.
Plants may form mutualisms with potential enemies as well. A limiting step in the nitrogen cycle is the capture of inorganic nitrogen from the air and its incorporation into organic forms plants can use. Bacteria have developed this ability, called nitrogen fixation, and are a critical link in this cycle. Legumes and some other plants have formed associations with certain bacteria, particularly the genus Rhizobium, in which the bacteria live in swellings, or nodules, on the plant roots. But since many bacteria are enemies (pathogens) of plants, plants and Rhizobium have had to reach a coevolved accommodation. Through coevolution, they have developed a specialized interaction that depends on manipulating expression of each other's genes. Rhizobium produces signals that turn off plant defense responses and identify it as friendly to the plant. Host plants produce chemical signals that turn on genes in the bacteria that produce signals directing the plant root to produce a nodule. The bacteria then invade the nodule, where the plant provides necessary nutrients in return for nitrogen. It is clear that this relationship has evolved from a battle between enemies, host and pathogen, to a mutualism.
Scientists are divided about how many species have been shaped by coevolution. Several important questions need to be answered before this issue will be resolved. If insects exert relatively little pressure on plants, how often would plant defenses change? Do insects make mistakes in selecting plants as food or oviposition sites? If not, how do they ever begin feeding on a new plant type? How great a change is necessary to provoke a response in the coevolutionary partner; for example, how much change in the shape of an orchid is necessary to provide improved visitation by an insect? And how can we evaluate the importance of the coevolutionary partner versus other factors that influence the evolution of plants, animals, and microbes?
see also Evolution of Plants; Interactions, Plant-Fungal; Interactions, Plant-Plant; Interactions, Plant-Vertebrate; Pollination Biology.
Jack C. Schultz
Price, Peter W. Insect Ecology. New York: John Wiley & Sons, Inc. 1997.
Schoonhoven, L. M., T. Jermy, and J. J. A. van Loon. Insect-Plant Biology. London: Chapman and Hall, 1998.
Thompson, John N. The Coevolutionary Process. Chicago, IL: University of Chicago Press, 1994.
——, and John J. Burdon. "Gene-for-Gene Coevolution Between Plants and Parasites." Nature 360 (1992): 121-25.
FIGS AND FIG WASPS
More than nine hundred species of figs (Ficus ) are pollinated by figs wasps (family Agaonidae) in rela tionships that exhibit closely coe volved characteristics. The hollow fig inflorescence is formed by a swollen flower receptacle (base) and is lined inside with flowers that go through five stages:
- Prefemale , in which the fig is closed to wasps as flowers de velop;
- Female , in which tiny wasps crawl inside the inflorescence through a special pore and lay eggs in the mature flowers;
- Interfloral , during which wasp lar vae develop inside some female flowers while others produce seed;
- Male , in which male flowers ma ture, producing pollen, while the new generation of wasps emerges from female flowers. Female wasps mate, collect pollen, and exit through escape holes bored by males; and
- Postfloral , in which seeds ripen, and the fruit becomes attractive to animals that disperse it.
The escaped females invade new fig flowers on other trees, repeating the cycle. The figs provide special ized flowers in which the wasps lay eggs, sacrificing these as a reward, and the timing of male and female flower production is designed to match the wasps' development. The wasps are specifically adapted for life in the fig, and cannot lay eggs or feed anywhere else. Usually only one wasp species can live in one fig species. A natural consequence of this system is that the figs we eat contain some of the minute wasps that do not escape.
Species are said to "coevolve" when their respective levels of fitness depend not only on their own genetic structure and adaptations but also the development of another species as well. The gene pool of one species creates selection pressure for another species. Although the changes are generally reciprocal, they may also be unilateral and still be considered coevolutionary.
The process of coevolution arises from interactions that establish structure in communities. A variety of different types of interactions can occur–symbiotic, where neither member suffers, or parasitic, predatory, and competitive relationships, where one member of a species pair suffers.
Coevolution can result from mutually positive selection pressure. For example, certain plants have in an evolutionary sense created positive situations for insects by providing valuable food sources for them. In return the insects provide a means to distribute pollen that is more efficient than the distribution of pollen by wind. Unfortunately, the plant and the insect species could evolve into a position of total dependency through increased specialization, thus enhancing the risk of extinction if either species declines.
Coevolution can also arise from negative pressures. Prey species will continually adapt defensive or evasive systems to avoid predation. Predators respond by developing mechanisms to surmount these defenses. However, these species pairs are "familiar" with one another, and neither of the strategies is perfect. Some prey are always more vulnerable, and some predators are less efficient due to the nature of variability in natural populations. Therefore the likelihood of extinction from this association is limited.
Several factors influence the likelihood and strength of coevolved relationships. Coevolution is more likely to take place in pairs of species where high levels of co-occurrence are present. It is also common in cases where selective pressure is strong, influencing important functions such as reproduction or mortality . The type of relationship—be it mutualistic, predator-prey, or competitor—also influences coevolution. Species that have intimate relationships, such as that of a specialist predator or a host-specific parasite, interact actively and thus are more likely to influence each other's selection. Species that do not directly encounter each other but interact through competition for resources are less likely candidates to coevolve, but the strength of the competition may influence the situation.
The result of coevolved relationships is structure in communities. Coevolution and symbiosis create fairly distinct relationships in communities, relationships that are mirrored in distribution, community energetics, and resistance to disturbances in species. Coevolution allows for recurring groupings of plant and animal communities.
[David A. Duffus ]
Barth, F. G. Insects and Flowers: The Biology of a Partnership. Princeton, NJ: Princeton University Press, 1991.
Erickson, J. The Living Earth: The Coevolution of the Planet and Life. Blue Ridge Summit, PA: Tab Books, 1989.