Chapter13 Competition

Along a coral reef off the north coast of Jamaica, threespot damselfish guard small territories of less than 1 m² (fig. 13.1 ). These small territories are regularly dispersed across the reef and contain most of the resources upon which the damselfish depend: nooks and crannies for shelter against predators, a carefully tended patch of fast-growing algae for food, and in the territories of males, an area of coral rabble kept clean for spawning. The damselfish constantly patrol and survey the borders of their territories, vigorously attacking any intruder that presents a threat to their eggs and developing larvae, or to their food supply. If you look carefully, however, you may find that not all members of the population have a territory. Damselfish without territories live in marginal areas around the territorial members, wandering from one part of the reef to another.

FIGURE 13.1 Territorial reef fish, such as this threespot damselfish, Eupomacentrus planifrons, compete intensely for space.

If you create a vacancy on the reef by removing one of the damselfish holding a territory, other damselfish appear within minutes to claim the vacant territory. Some of the new arrivals are threespot damselfish like the original resident, and some are cocoa damselfish, which generally live a bit higher on the reef face. These new arrivals fight fiercely for the vacated territory. The damselfish chase each other, nip each other’s flanks, and slap each other with their tails. The melee ends within minutes, and life among the damselfish settles back into a kind of tense tranquillity. The new resident, which may have driven off a half dozen rivals, is usually another threespot damselfish.

This example demonstrates several things. First, individual damselfish maintain possession of their territories through ongoing competition with other damselfish, and this competition takes the form of interference competition, which involves direct aggressive interaction between individuals. Second, though it may not appear so to the casual observer, there is a limited supply of suitable space for damsel fish territories, a condition that ecologists call resource limitation. Third, the threespot damselfish are subject to intraspecific competition, competition with members of their own species, as well as nterspecific competition, competition between individuals of two species that reduces the fitness of both. The effects of competition on the two competitors may not be equal, however The individuals of one species may suffer greatly reduced fitness while those of the second are affected very little. The observation that threespots generally win in aggressive encounters with cocoa damselfish suggests this sort of competitive asymmetry.

Competition is not always as dramatic as fighting damselfish nor is it always resolved so quickly-. In a mature white pine forest in New Hampshire, tree roots grow throughout the soil taking up nutrients and water as they provide support. In 1931, J. Toumey and R. Kienholz designed an experiment to determine whether the activities of these tree roots suppress the activities of other plants. The researchers cut a trench, 0.92 m deep, around a plot 2.74 m by 2.74 m in the middle of the forest. In so doing, they cut 825 roots, which removed potential competition by these roots for soil resources. They also established control plots on either side of the trenched plot and then watched as the results of their experiment unfolded. The experiment continued for 8 years, with retrenching every 2 years and over 100 roots cut each time. By retrenching, the researchers maintained their experimental treatment, suppression of potential root competition.

In the end, this 8-year experiment yielded results as dramatic as those with the damselfish. Vegetative cover on the section of forest floor that had been released from root competition was 10 times that present on the control plots. Apparently the roots of white pines exert interspecific competition for some combination of nutrients and water that is strong enough to suppress the growth of forest floor vegetation (fig. 13.2). In addition, the growth of young white pines was also much greater within the trenched plots than in the control plots. Therefore, considerable intraspecific competition also occurred on the forest floor.

FIGURE 13.2 Competition in a forest can be as intense as competition on a coral reef However, much of the competition in a forest takes place underground where the roots of plants compete for water and nutrients.

Ecologists have long thought that both interspecific and intraspecific competition are pervasive in nature. For instance, Darwin thought that interspecific competition was an important source of natural selection. While ecologists have shown that interspecific competition substantially influences the distribution and abundance of many species, they have also questioned the assumption that competition is an all-important organizer of nature. Such questioning has stimulated more careful research and more rigorous testing of the influence of competition on populations, and while this testing continues, sufficient evidence has accumulated to make some tentative generalizations.

CONCEPTS

l Studies of intraspecific competition provide evidence for resource limitation.

l The niche reflects the environmental requirements of species.

l Mathematical and laboratory models provide a theoretical foundation for studying competitive interactions in nature.

l Competition can have significant ecological and evolutionary influences on the niches of species.

CASE HISTORIES: resource competition

Studies of intraspecific competition provide evidence for resource limitation.

In chapter 11, we saw that slowing population growth at high densities produces a sigmoidal, or S-shaped, pattern in which population size levels off at carrying capacity. Our assumption in that discussion was that intraspecific competition for limited resources plays a key role in slowing population growth at higher densities. The effect of intraspecific competition is included in the model of logistic population growth. If competition is an important and common phenomenon in nature, then we should be able to observe it among individuals of the same species, individuals with identical or very similar resource requirements. Thus we begin our discussion of competition with intraspecific competition.

Intraspecific Competition Among Herbaceous Plants

In chapter 6, we reviewed experiments by David Tilman and M. Cowan (1989) that showed how plants alter root: shoot ratios in response to availability of soil nitrogen. The plants in these experiments reduced their allocation to roots as soil nitrogen concentration increased. The experiments also included evidence for intraspecific competition. Tilman and Cowan grew the grass Sorghastrum nutans at low density (7 plants per pot) and high density (100 plants per pot). The results showed that the root: shoot ratios are higher when the plants are grown at high density, suggesting that competition for nutrients was more intense under

these conditions.

The results of Tilman and Cowan's experiments also show that soil nitrogen concentration and population density substantially influence growth rates and individual plant weight. For example, the weight of S. nutans increased with increased soil nitrogen (fig. 13.3). Therefore, we can conclude that both these responses were limited by nitrogen availability at the lower concentrations in the experiment. Now compare the growth rates and plant weights shown by plants grown at low and high densities. How are they different? Both growth rate and plant weight are higher in the low-density populations, and we can conclude that competition for nutrients (resources) is more intense at the higher plant population density. Such competition for limited resources in natural populations usually leads to mortality among the competing plants.

Self-Thinning in Plant Populations

The development of a stand of plants from the seedling stage to mature individuals suggests competition for limited resources. Each spring as the seeds of annual plants geminate, their population density often numbers in the thousands

FIGURE 13.3 Population density, soil nitrogen, and the size attained by the grass Sorghastrum nutans (data from Tilman and Cowan 1989).

per square meter. However, as the season progresses and individual plants grow, population density declines. This same pattern occurs in the development of a stand of trees. As the stand of trees develops, more and more biomass is composed of fewer and fewer individuals. This process is called self. thinning.

Self-thinning appears to result from intraspecific competition for limited resources. As a local population of plants develops, individual plants take up increasing quantities of nutrients, water, and space for which some individuals compete more successfully. The losers in this competition for resources die, and population density decreases, or "thins," as a consequence. Over time the population is composed of fewer and fewer large individuals.

One way to represent the self-thinning process is to plot total plant biomass against population density. If we plot the logarithm of plant biomass against the logarithm of plant density, the slope of the resulting line averages around –¹/₂. In other words there is an approximately one-unit increase in total plant biomass with each two-unit decrease in population density; plant population density more rapidly than biomass increases (fig. 13.4).

FIGURE 13.4 Self-thinning in plant populations (data from Westoby 1984).

Another way to represent the self-thinning process is to plot the average weight of individual plants in a stand against density (fig. 13.5). The slope of the line in such plots averages around -³/₂. Because self-thinning by many species of plants comes close to a -³/₂ relationship, this relationship has come to be called the -³/₂ self-thinning rule.The -³/₂ self-thinning rule was first proposed by K. Yoda and colleagues (1963) and amplified by White and Harper (1970), who provided many additional examples (e.g., fig 13.5). Subsequently, the self-thinning .rule became widely accepted among ecologists.

FIGURE 13.5 Self-thinning in populations of alfalfa, Medicago sativa (data from White and Harper 1970).

Recent analyses have shown that self-thinning in some plant populations deviates significantly from the -³/₂ (or -¹/₂ for biomass-numbers) slope. However, regardless of the precise trajectory followed by different plant populations, self-thinning of plant populations has been demonstrated repeatedly. The important point, from the perspective of our present discussion, is that self-thinning occurs and appears to be the consequence of intraspecific competition for limited resources. Resource limitation has also been demonstrated in experiments on intraspecific competition within animal populations.

Intraspecific Competition Among Planthoppers

Ecologists have often failed to demonstrate that insects, particularly herbivorous insects, compete. However, one group of insects in which competition has been repeatedly demonstrated are the Homoptera, including the leafhoppers, planthoppers, and aphids. Robert Denno and George Roderick (1992), who studied interactions among planthoppers (Homoptera, Delphacidae), attribute the prevalence of competition among the Homoptera to their habit of aggregating, to rapid population growth, and to the mobile nature of their food supply, plant fluids.

Denno and Roderick demonstrated intraspecific competition within populations of the planthopper Prokelesisia marginata, which lives on the salt marsh grass Spartina alternifiora along the Atlantic and Gulf coasts of the United States. The population density of P. marginata was controlled by enclosing the insects with Spartina seedlings at densities of 3, 11, and 40 leafhoppers per cage, densities that are within the range at which they live in nature. At the highest density, P. marginata showed reduced survivorship, decreased body length, and increased developmental time (fig. 13.6). These signs of intraspecific competition were probably the result of reduced food quality at high leafhopper densities. Plants heavily populated by planthoppers show reduced concentrations of protein, chlorophyll, and moisture. Therefore, competition between these leafhoppers was probably the result of limited resource supplies. However, as demonstrated in the following example, intraspecific interference competition may occur in the absence of obvious

resource limitation.

FIGURE 13.6 Population density and planthopper performance (data from Denno and Roderick 1992).

Interference Competition Among Terrestrial Isopods

Edwin Grosholz (1992) used a field experiment to study the effects of a wide range of biotic interactions on the population biology of the terrestrial isopod Porcellio scabet: This organism, which is associated with human activities such as farming and gardening and is found throughout the world, sometimes lives at densities in excess of 2,000 individuals per square meter. Such high densities suggest a strong potential for intraspecific competition.

Grosholz conducted his experiments on an outdoor grid of 48, 0.36 m2 plots enclosed by aluminum flashing. To control isopod movements, he buried the flashing 12.5 cm into the soil and extended it 12.5 cm above the soil surface. Two experimental treatments were used: (1) to test for food limitation, the food within the enclosures was supplemented by adding sliced carrots and potatoes, and (2) to test for density effects, study plots were stocked with either 100 ox 50 P. scaber Supplementing food had no effect on survival by P scaber. However, survival was lower at the higher population density (fig. 13.7). Crcosholz attributed lower survival at the higher density to cannibalism, a common occurrence in terrestrial isopods.

FIGURE 13.7 Population density and survival in populations of a terrestrial isopod, Porcellio scaler (data from Grosholz 1992).

Do you think increasing the population densities in the experiment might have changed the results? Since densities in nature sometimes exceed 2,000 per square meter, food limitation might not be observed until such densities are approached. How might the interpretation of the experiment have been altered if other indicators of competition, such as growth rate, size, and reproductive rate, were measured? While food supplements did not affect survival, these other unmeasured attributes may have been affected. Despite these limitations, the study offers interesting insights into the role that interference may play in intraspecific competition, even

in the absence of obvious resource limitation.

As we move from discussions of intraspecific to interspecific competition, we need to back up a bit and consider how we might portray the environmental requirements of species. We do this because interspecific competition usually occurs among species with similar environmental requirements, that is, among species with similar niches.

CASE HISTORIES: niches

The niche reflects the environmental requirements of species.

The word niche has been in use a long time. Its earliest and most basic meaning was that of a recessed place in a wall where one could set or display items. For about a century, however, ecologists have given a broader meaning to the word. To the ecologist, the niche summarizes the environmental factors that influence the growth, survival, and reproduction of a species. In other words, a species' niche consists of all the factors necessary for its existence--approximately

when, where, and how a species makes its living.

The niche concept was developed independently by Joseph Grinnell (1917, 1924) and Charles Elton (1927), who used the term niche in slightly different ways. In his early writings, Grinnell's ideas of the niche centered around the influences of the physical environment, while Elton's earliest concept included biological interactions as well as abiotic factors. However their thinking and emphasis may have differed, it is clear that the views of these two researchers had much in common and that our present concept of the niche rests squarely on their pioneering work.

The niche concept was developed over a period of several decades; however, it was within the context of interspecific competition that the importance of the niche concept was fully realized. It was the work of G. F. Ganse (1934), whose principal interest was interspecific competition, that ensured a prominent place for the niche concept in modem ecology. Particularly important was Gause's competitive exclusion principle, which states that two species with identical niches cannot coexist indefinitely. Gause experimented with competition in the laboratory and obtained results indicating that when two species compete, one will be a more effective competitor for limited resources, that is, will be more effective at convetting resources into offspring. As a consequence, the more effective competitor will have higher fitness (higher reproductive success) and eventually excludes all individuals of the second species. The competitive exclusion principle set the niche concept in a broader context. After Gause, describing the niches of species was no longer an end in itself but a steppingstone to understanding interactions between species---a potential key to understanding the organization of nature.

Though the work of Gause played a central role in the development of the niche concept, a rigorous definition of the niche awaited later ecologists. We can now point to a single paper authored by G. Evelyn Hutchinson (1957) as the agent that crystallized the niche concept and stimulated the work of an entire generation of ecologists. In this seminal paper titled simply, "Concluding Remarks," Hutchinson defined the niche as an n-dimensional hypervolume, where n equals the number of environmental factors important to the survival and reproduction by a species. Hutchinson called this hypervolume, which specifies the values of the n environmental factors permitting a species to survive and reproduce, as the fundamental niche of the species. The fundamental niche defines the physical conditions under which a species might live, in the absence of interactions with other species. However, Hutchinson recognized that interactions such as competition may restrict the environments in which a species may live and referred to these more restricted conditions as the realized niche. While Hutchinson was particularly concerned with the influence of competition on the realized niche, later authors have pointed out that other interactions such as predation, disease, and parasitism may also be important in restricting the distribution of species.

In a single word, niche captures most of what we discussed in sections II and HI, where we considered how environment affects the growth, survival, reproduction, distribution, and abundance of species. So, why introduce the niche concept here? The reason is that we, like the first ecologists to use the term, need a concept that represents all the environmental requirements of a species. The niche concept carries us beyond the details of individual species' requirements to a position where we can more easily consider the ecology of interactions between species, interactions such as competition, predation, and mutualism.

Do you think it's possible to completely describe Hutchinson's n-dimensional hypervolume niche for any species? Probably not, since there are so many environmental factors that potentially influence survival and reproduction. Fortunately, it appears that niches are often determined mostly by a few environmental factors and so ecologists are able to apply a simplified version of Hutchinson's comprehensive niche concept. In studies of animals, ecologists have frequently described niches in terms of their feeding biology.

The Feeding Niches of Galapagos Finches

As we saw in chapter 10, availability of suitable food significantly affects the survival and reproduction of Galapagos finches. In other words, food has a major influence on the niches of Galapagos finches. Because the kinds of food used by birds is largely reflected by the form of their beaks, Peter Grant (1986) and his colleagues were able to represent the feeding niches of Galapagos finches by measuring their beak morphology. For instance, differences in beak size among small, medium, and large ground finches translate directly into differences in diet. The large ground finch, Geospiza magnirostris, eats larger seeds; the medium ground finch, G. fortis, eats medium-sized seeds; while the small ground finch, G. fuliginosa, eats small seeds (fig. 13.8).

FIGURE 13.8 Relationship between body size and seed size in Galapagos finch species (data from Grant 1986).

The size of seeds that can be eaten by Galapagos finches can be estimated by simply measuring the depths of their beaks. Studies of seed use by G. fortis on Daphne Major showed clearly that even within species, beak size affects the composition of the diet. Within this population, individuals with the deepest beaks fed on the hardest seeds, while individuals with the smallest beaks fed on the softest seeds (fig. 13.9).

FIGURE 13.9 Relationship between the hardness of seeds eaten by medium ground finches, Geospiza fortis, and beak depth (data from Boag and Grant 1984).

The importance of beak size to seed use was also demonstrated by the effects of the 1977 drought on the G. fortis population of Daphne Major. In chapter 11, we saw how this drought caused substantial mortality in this population (see fig. 11.16). However, this mortality did not fall equally on all segments of the population. As seeds were depleted, the birds ate the smallest and softest seeds first, leaving the largest and toughest seeds (fig. 13.10). In other words, following the drought not only were seeds in short supply, the remaining seeds were also tougher to crack. Because they could not crack the remaining seeds, mortality fell most heavily on smaller birds with smaller beaks. Consequently, at the end of the drought, the G. fortis population on Daphne Major was dominated by larger individuals that had survived by feeding on hard seeds (fig. 13.11).

FIGURE 13.10 Seed depletion by the medium ground finch, Geospiza fortis, and average seed hardness (data from Grant 1986).

FIGURE 13.11 Selection for larger size among medium ground finches, Geospiza fortis, during a drought on the island of Daphne Major (data from Grant 1986)

These studies show that beak size provides significant insights into the feeding biology of Galapagos ground finches. Since food is the major determinant of survival and reproduction among these birds, beak morphology gives us a very good picture of their niches. However, the niches of other kinds of organisms are determined by entirely different environmental factors. Let's consider the niche of a dominant species in salt marshes.

The Habitat Niche of a Salt Marsh Grass

Biologists discovered Spartina anglica approximately one century ago, as a new species recently produced by allopolyploidy (fig. 13.12). Allopolyploidy is a process of speciation initiated

FIGURE 13.12 The salt marsh grass Spartina anglica.

by hybridization of two different species. S. anglica arose initially as a cross between S. mar/t/ma, a European species, and S. alternifiora, a North American species. At least one of these hybrid plants later doubled its chromosome number, making it capable of sexual reproduction, and produced a new species: S. anglica. From its center of origin in Lymington, Hampshire, England, S. anglica spread northward along the coasts of the British Isles. During this same period, it colonized the coast of France and was widely planted elsewhere in northwest Europe as well as along the coasts of New Zealand, Australia, and China. The Chinese population of this salt marsh grass, established from only 21 plants in 1963, grew to cover 36,000 ha by 1980. S. anglica is extensively planted for stabilizing mudflats because it is more tolerant of periodic inundation and water-sat-urated soils than most other salt marsh plants. This environmental tolerance is reflected in the distribution of the plant in northwestern Europe, where it generally inhabits the most seaward zone of any of the salt marsh plants.

The local distribution of S. anglica in the British Isles is well predicted by a few physical variables related to the duration and frequency of inundation by tides and waves. The lower and upper intertidal limits of the grass are mainly determined by the magnitude of tidal fluctuations during spring tides. Where tidal fluctuations are greater, both the lower and upper limits are higher on the shore. However, throughout its British range, the gross generally occupies the intertidal between mean high-water spring tides and mean high-water neap tides (fig. 13.13). A second factor that determines the local distribution of S. anglica is the fetch of the estuary. The fetch of a body of water is the longest distance over which wind can blow and is directly related to the maximum size of waves that can be generated by wind. All other factors being equal, larger waves occur on estuaries with greater fetch. The larger the fetch the higher S. anglica must live in an estuary to avoid disturbance by waves.

FIGURE 13.13 The niche of Spartina anglica is related to tidal fluctuations.

The upper limit of S. anglica's distribution within the intertidal zone is also negatively correlated with latitude. In northerly locations within the British Isles, the grass does not occur quite as high in the intertidal zone as it does in the south. What factors might restrict the distribution at northern sites? One factor we should consider is that & anglica is a C₄ plant. Remember from chapter 6 that C₄ grasses generally do better in warm environments. In northerly locations, S. anglica is replaced in the upper intertidal zone by C₃ plants. Could it be that competition with these C₃ plants at northern sites excludes S. anglica from the upper intertidal zone? We'll take up this question later in the chapter when we discuss experimental approaches to the study of competition.

CASE HISTORIES: mathematical and laboratory models

Mathematical and laboratory models provide a theoretical foundation for studying competitive interactions in nature.

As ecologists have used models to explore the ecology of competition, mathematical and laboratory models have played complementary roles. Both mathematical and laboratory models are generally much simpler than the natural circumstances the ecologist wishes to understand. However, while sacrificing accuracy, this simplicity offers a degree of control that ecologists would not have in most natural settings.

D. B. Mertz (1972) began a review of four decades of research on Tribolium beetle populations with an astute summary of the characteristics of models in general and of the "Tribolium model" in particular: (1) It is an abstraction and simplification, not a facsimile, of nature; (2) except for the beetles themselves, it is a man-made construct, partly empirical and partly deductive; and (3) it is used to provide insights into natural phenomena. The predictions of these simplified models can be tested in natural systems and either supported or falsified. If falsified, a theory can be modified to accommodate the new information. Ideally, scientific understanding proceeds as a consequence of this dialog between theory and observation, between theoretician and empiricist.

Modeling Interspecific Competition

As we saw in chapter 11, the model of logistic population growth includes a term for intraspecific competition but can be expanded to include the influence of interspecific competition on population growth. The first to do so was Vito Volterra (1926), who was interested in developing a theoretical basis for explaining changes in the composition of a marine fish community in response to reduced fishing during World War I. Alfred Lotka (1932) independently repeated Volterra's analysis and extended it using graphics to represent changes in the population densities of competing species during competition.

Let's retrace the steps of Lotka's and Volterra's modeling exercise, beginning with the logistic model for population growth discussed in chapter 11:

dN / dt = r_m N [ (K–N) / K ]

We can express the population growth of two species of potential competitors with the logistic equation:

dN₁ / dt = r_m1 N₁ [ (K₁–N₁) / K₁ ] and dN₂ / dt = r_m2 N₂ [ (K₂–N₂) / K₂ ]

Where N₁ and N₂ are the population sizes of species 1 and 2, K₁ and K₂ are their carrying capacities, and r_m1 and r_m2 are the intrinsic rates of increase for species 1 and 2. In these models, population growth slows as N increases and the relative level of intraspecific competition is expressed as the ratio of numbers to carrying capacity, either N₁/K₁ or N₂/K₂. The assumption here is that resource supplies will diminish as population size increases due to intraspecific competition for resources. Resource levels can also be reduced by interspecific competition.

Lotka and Volterra included the effect of interspecific competition on the population growth of each species as:

dN₁ / dt = r_m1 N₁ [ (K₁–N₁ –α₁₂N₂) / K₁ ]

and

dN₂ / dt = r_m2 N₂ [ (K₂–N₂ –α₂₁N₁ ) / K₂ ]

In these models, the rate of population growth of a species is reduced both by conspecifics (individuals of the same species) and by individuals of the competing species, that is, interspecific competition. The effects of intraspecific competition (- N _l and – N ₂) are already included in the

logistic models for population growth. The effect of interspecific competition is incorporated into the Lotka-Volterra model by –α₁₂N₂ and –α₂₁N_1. The terms α₁₂ and α₂₁ are called competition coefficients and express the competitive effects of the competing species. Specifically, α₁₂ is the effect of an individual of species 2 on the rate of population growth of species 1, while α₂₁ is the effect of an individual of species 1 on the rate of population growth of species 2. In this model, interspecific competitive effects are expressed in terms of intraspecific equivalents. If, for example, α₁₂ > 1, then the competitive effect of an individual of species 2 on the population growth of species 1 is greater than that of an individual of species 1. If, on the other hand, α₁₂ < 1, then the competitive effect of an individual of species 2 on the population growth of species 1 is less than that of an individual of species 1.

In general, the Lotka-Volterra model predicts coexistence of two species when, for both species, interspecific competition is weaker than intraspecific competition. Otherwise, one species is predicted to eventually exclude the other. These conclusions come from the following analysis.

Populations of species 1 and 2 stop growing when:

dN₁ / dt = r_m1 N₁ [ (K₁–N₁ –α₁₂N₂) / K₁ ] = 0

and

dN₂ / dt = r_m2 N₂ [ (K₂–N₂ –α₂₁N₁ ) / K₂ ] = 0

That is, when:

(K₁–N₁ –α₁₂N₂) = 0 and (K₂–N₂ –α₂₁N₁ ) = 0

Or, rearranging these equations, we predict that population growth for the two species will stop when:

N₁ = K₁ –α₁₂N₂ and N₂ = K₂–α₂₁N₁

These are equations for straight lines, called isoclines of zero population growth, where everywhere along the lines population growth is stopped:

dN₁ / dt = 0 and dN₂ / dt = 0

Above an isocline of zero growth, the population of a species is decreasing; below it the population is increasing (fig. 13.14).

FIGURE 13.14 The orientation of isoclines for zero population growth and the outcome of competition according to the Lotka-Volterra competition model.

The isoclines of zero growth show how the environment can be filled up or, in other words, the relative population sizes of species 1 and species 2 that will deplete the critical resources. At one extreme, for example, for species 1, the environment is completely filled by species 1 and species 2 is absent. This occurs where N₁ = K₁. At the other extreme, again for species 1, the environment can be saturated entirely by species 2, while species 1 is absent. This occurs where N₂ = K₁/α₁₂. In between these extremes, the environment is saturated with a mixture of species 1 and 2. The graph of the isocline for zero growth for species 2 can be interpreted in a similar way.

Putting the isoclines of zero growth for the two species on the same axis allows us to predict if one species will exclude the other or whether the two species will coexist. The precise prediction depends upon the relative orientation of the two isoclines. As shown in figure 13.14, there are four possibilities.

The Lotka-Volterra model predicts that one species will exclude the other when the isoclines do not cross. If the isocline for species I lies above that of species 2, species 1 will eventually exclude species 2. This exclusion occurs because all growth trajectories lead to the point where N₁ = K₁ and N₂ = 0 (fig. 13.14a). Figure 13.14b portrays the opposite situation in which the isecline for species 2 lies completely above that of species 1 and species 2 excludes species I. In this case, all trajectories of population growth lead to the point where N₂ = K₂ and N₁ = 0.

Coexistence is possible only in the situations in which the iseclines cross. However, only one of these situations leads to stable coexistence. Figure 13.14c shows the situation in which coexistence is possible at the point where the isoclines of zero population growth cross but coexistence is unstable. In this situation, K₁ > K₂/α₂₁ and K₂ > K₁/α₁₂ and most population growth trajectories lead either to the points where N₁ = K₁ and N₂ = 0, or to where N₂ = K₂ and N₁ = 0. The populations of species 1 and 2 may arrive at the point where the lines cross, but any environmental variation that moves the populations off this point eventually leads to exclusion of one species by the other. Figure 13.14d represents the only situation that predicts stable coexistence of the two species. In this situation, K₂/α₂₁ > K₁ and K₁/α₂₁ > K₂ and all growth trajectories lead to the point where the isoclines of zero growth cross.

What is the biological meaning of saying that all growth trajectories lead to the point where the isoclines of zero growth cross? What this means is that the relative abundances of species 1 and 2 will eventually arrive at the point where the isoclines cross, a point where the abundances of both species are greater than zero. In this situation, each species is limited more by members of their own species than they are by members of the other species. In other words, the Lotka-Volterra model predicts that species coexist when intraspecific competition is stronger than interspecific competition. This prediction is supported by the results of laboratory experiments on interspecific competition.

Laboratory Models of Competition

Experiments with Paramecia

G. F. Gause (1934) used laboratory experiments to test the major predictions of the Lotka-Volterra competition model. During the course of his work Gause experimented with many organisms, but the most well known of his experimental subjects were paramecia. Paramecia are freshwater, ciliated protozoans that offer several advantages for laboratory work. First, since they are small, they can be kept in large numbers in a small space and some of their natural habitats are fairly well simulated by laboratory aquaria. In addition, paramecia feed on microorganisms, which can be cultured in the laboratory and provided in whatever concentration desired by the experimenter.

In one of his most famous experiments, Ganse studied competition between Paramecium caudatum and P. aurelia. The question he posed was: Would one of these two species drive the other to extinction if grown together in microcosms where they were forced to compete with each other for a limited food supply?

Ganse demonstrated resource limitation by growing pure populations of P. caudatum and P. aurelia in the presence of two different concentrations of their food, the bacterium Bacillus pyocyaneus. If food supplies limit the growth of laboratory populations of these paramecia, what kind of population growth would you expect them to show? As you probably expect, Ganse observed sigmoidal growth with a clear carrying capacity at both full- and half-strength con-

centrations of the food supply (fig. 13.15). When grown in the presence of a full-strength concentration of food, the carrying capacity of P aurelia was 195. When food availability was halved, the carrying capacity of this species was reduced to 105. P. caudatum showed a similar response to food concentration. In the presence of a full-strength concentration of food, P. caudatum had a carrying capacity of 137. At a half-strength concentration, the carrying capacity was 64. The nearly one-to-one correspondence between food level and the carrying capacities of these two species provides evidence that when grown alone, the carrying capacity was determined by intraspecific competition for food. These results set the stage for Gause's experiment to determine whether interspecific competition for food, the limiting resource in this system, would lead to the exclusion of one of the competing species.

FIGURE 13.15 Population growth and population sizes attained by Paramecium aurelia and P. caudatum grown separately (data from Gause 1934).

When grown together, P. aurelia survived, while the population of P. caudatum quickly declined. The difference in results obtained at the two food concentrations support the conclusion that competitive exclusion results from competition for food. At a full-strength food concentration, the decline in the P. caudatum population was approaching exclusion by 16 days but exclusion was not complete. In contrast, at a half-strength food concentration, P caudatum had been entirely eliminated by day 16. What does this contrast in the time to exclusion suggest about the influence of food supply on competition? It suggests that reduced resource supplies increase the intensity of competition.

Experiments with Flour Beetles

Tribolium, beetles of the family Tenebrionidae, infest stored grains and grain products. The discovery of an infestation of these beetles in an urn of milled grain in the tomb of an Egyptian pharaoh buried about 4,500 years ago suggests that these beetles have been engaged in this occupation for some time. Their habit of attacking stored grains makes them a convenient laboratory model. Since all life stages of Tribolium live in finely milled flour, small containers of flour provide all the environmental requirements to sustain a population. R. N. Chapman (1928) began working with laboratory populations of Tribolium at the University of Chicago in the 1920s, where ever since, work has focused on two species: T. confusum and T. castaneum.

Thomas Park (1954) worked extensively on interspecific competition between these two species under six environmental conditions: hot-wet (34℃, 70% RH, relative humidity), hot-dry (34℃, 30% RH), temperate-wet (29℃, 70% RH), temperate-dry (29℃, 30% RH), cool-wet (24℃, 70% RH), and cool-dry (24℃, 30% RH). In environments held at 34℃ and 70% relative humidity, both species established healthy populations that persisted over the entire duration of the experiment

(fig. 13.16a). However, when grown together under these conditions, T. castaneurn usually excludes T. confusum (fig. 13.16b). Cool-dry conditions appear to favor T. confusum. Even when grown by itself at 24℃ and 30% relative humidity, T. castaneum does not persist for long (fig. 13.17a). This species quickly disappears from mixed cultures held at 24℃ and 30% relative humidity (fig. 13.17b).

FIGURE 13.16 Populations of Tribolium confusum and T castaneum grown separately (a) and together (b) at 34℃ and 70% relative humidity (data from Park 1954).

FIGURE 15.17 Populations of Tribolium confusum and T castaneum grown separately (a) and together (b) at 24℃ and 30% relative humidity (data from Park 1954).

Under intermediate environmental conditions each species did well when grown alone but the outcome of interspecific competition was not completely predictable. T. castaneum won 86% of the trials under temperate-wet conditions, while T. confusum won 71% to 90% of the trials under hot-dry, temperate-dry, and cool-wet conditions. Notice that under a particular set of conditions either T castaneum or T. confusum was usually favored, but not always.

How can we interpret the results of these laboratory experiments in terms of the effects of competition on these species' niches? Growing the two species separately showed that the fundamental niche of T. castaneum includes five of the six environmental conditions in the experiment, while the fundamental niche of T. confusum includes all six environmental conditions. Growing the two species together suggests that interspecific competition restricts the realized niches of both species to fewer environmental conditions. Based upon these results, can we conclude that interspecific competition restricts the realized niches of species in nature? No, this would be going beyond the proper role for a model system, which is best used to generate hypotheses and guide experimentation on natural populations.

CASE HISTORIES: competition and niches

Competition can have significant ecological and evolutionary influences on the niches of species.

Competition can have short-term ecological effects on the niches of species by restricting them to realized niches. These species may retain their capacity to inhabit the fuller range of environments we call the fundamental niche. However, if competitive interactions are strong and pervasive enough, they may produce an evolutionary response in the competitor population--an evolutionary response that changes the fundamental niche. In this section, we explore the evidence for both ecological and evolutionary influences on the niches of natural populations. Field experiments show that interspecific competition may restrict the niches of populations in nature.

Niches and Competition Among Plants

A. Tansley (1917) conducted one of the first experiments to test whether competition was responsible for the separation of two species of plants on different soil types. In the introduction to his paper, Tansley pointed out that while the separation of closely related plants had long been attributed to mutual competitive exclusion, it was necessary to perform manipulative experiments to demonstrate that this interpretation is correct. That is exactly what Tansley did to account for the mutually exclusive distributions of Galium saxatile and G. sylvestre (now G. pumilum), two species of small perennial plants commonly called bedstraw (fig. 13.18). In the British Isles, G. saxatile is largely confined to acidic soils and G. sylvestre to basic limestone soils.

FIGURE 13.18 These two species of bedstraw grow predominately on different soil types: Galium saxatile (shown here) grows mainly on acidic soils, while G. sylvestre (G. pumilum) grows mainly on basic soils.

Tansley conducted his experiment at the Cambridge Botanical Garden from .3911 to 1917, where seeds of the two species of plants were sown in planting boxes of acidic and basic soils. The seeds were sowed in single-species plantings and in mixtures of the two species. Both species germinated on both soil types, in both single- and mixed-species plantings (fig. 13.19). Like the paramecia studied by Gause, both Gallium species established healthy populations on both soil types when grown by themselves and these single-species plantings persisted to the end of the 6-year study. However, as the two species grew in mixed plantings, Tansley observed clear competitive dominance by each species on its normal soil type.

FIGURE 13.19 Percentage seed germination by Galium saxatile and G. sylvestre in basic calcareous soils and acidic peat soil (data from Tansley 1917).

On limestone soils, G. sylvestre, the species naturally found on limestone soils, overgrew and eliminated G. saxatile, the acidic soil species, by the end of the first growing season. On acidic soils, the relationship was reversed and G. saxatile was competitively dominant but competitive exclusion was not completed. Growth by both species was so slow on the acidic soils that it took until the end of the 6-year experiment for G. saxatile to completely cover the planting boxes containing acidic soils, a density attained by G. sylvestre on limestone soils in just I year. However, among the abundant G. saxatile Tansley found a few "quite healthy" plants of G. sylvestre. What do you think would have happened to the G. sylvestre on acidic soils if the experiment had been continued for a few more years? Of course it's impossible to say with certainty, but it is likely that G. saxatile would eventually exclude G. sylvestre. The delayed exclusion was probably due to the extremely slow growth of both species on acidic soils.

Tansley was one of the first ecologists to use experiments to demonstrate the influence of interspecific competition on the niches of species. The fundamental niche of both species of Galium included a wider variety of soil types than they inhabit in nature. The results of this experiment suggest that interspecific competition restricts the realized niche of each species to a narrower range of soil types.

Niche Overlap and Competition Between Barnacles

Like salt marsh plants, the barnacles Balanus balanoides and Chthamalus stellatus are restricted to predictable bands in the intertidal zone. We saw in chapter 9 (see fig. 9.8) that adult Chthamalus along the coast of Scotland are restricted to the upper intertidal zone, while adult Balanus are concentrated in the middle and lower intertidal zones. Joseph Connell's observations (I961) indicate that Balanus is limited to the middle and lower intertidal zones because it cannot withstand the longer exposure to air in the upper intertidal zone. However, physical factors only partially explain the distribution of Chthamalus. Connell noted that larval Chthamalus readily settle in the intertidal zone below where the species persists as adults but that these colonists die out within a relatively short period. In the course of field experiments, Connell discovered that interspecific competition with Balanus plays a key role in determining the lower limit of Chthamalus within the intertidal zone.

Because barnacles are sessile, small, and grow in high densities, they are ideal for field studies of survivorship. Their exposure at low tide is an additional convenience for the researcher. Connell established several study sites from the upper to the lower intertidal zones where he kept track of barnacle populations by periodically mapping the locations of every individual barnacle on glass plates. He established his study areas and made his initial maps in March and April of 1954, before the main settlement by Balanus in late April. He divided each of the study areas in half and kept one of the halves free of Balanus by scraping them off with a knife. Connell determined which half of each study site to keep Balanus-free by flipping a coin.

By periodically remapping the study sites, Connell was able to monitor interactions between the two species and the fates of individual barnacles. The results showed that in the middle intertidal zone Chthamalus survived at higher rates in the absence of Balanus (fig. 13.20). Balanus settled in densities up to 49 individuals per square centimeter in the middle intertidal zone and grew quickly, crowding out the second species in the process. In the upper intertidal zone, removing Balanus had no effect on survivorship by the second species because the population density of Balanus was too low to compete seriously. Connell's results provide direct evidence that Chthamalus is excluded from the middle intertidal zone by interspecific competition with Balanus.

FIGURE 13.20 A competition experiment with barnacles: removal of Balanus and survival by Chthamalus in the upper and middle intertidal zones (data from Connell 1961).

How does interspecific competition affect the niche of Chthamalus? In the absence of Balanus, it can live over a broad zone from the upper to the middle intertidal zones. Using the terminology of Hutchinson (1957), we can call this broad range of physical conditions the fundamental niche of Chthamalus. However, competition largely restricts Chthamalus to the upper intertidal zone, a more restricted range of physical conditions constituting the species' realized niche (fig. 13.21).

FIGURE 13.21 Environmental factors restricting the distribution of Chthamalus to the upper intertidal zone.

Does variation in interspecific competition completely explain the patterns seen by Connell? At the lowest levels in the lower intertidal zone, Chthamalus suffered high mortality even in the absence of Balanus (see fig. 13.20). What other factors might contribute to high rates of mortality by Chthamalus in the lower intertidal zone? Experiments have shown that this species can withstand periods of submergence of nearly 2 years, so it seems that it is not excluded by physical factors. It turns out that the presence of predators in the lower intertidal zone introduces complications that we will discuss in chapter 14 when we examine the influences of predators on prey populations.

Competition and the Habitat of a Salt Marsh Grass

How do you think competition might affect populations of the salt marsh grass Spartina anglica, whose niche we discussed earlier in the chapter? Field experiments have demonstrated that S. anglica, like Chthamalus, is restricted to its typical intertidal zone partly by interspecific competition with other salt marsh plants. In contrast to Chthamalus, however, S. anglica receives competitive pressure from the landward side of its intertidal distribution (Scholten and Rozema 1990, Scholten et al. 1987).

Does this reversal in the direction of competitive pressure make sense? It should, since in the case of barnacles we have marine organisms for which greater physical challenge occurs as they inhabit areas higher in the intertidal zone. In the case of the salt marsh plants, we are dealing with organisms descended from terrestrial ancestors that are met with increasing physical challenge as they inhabit areas lower in the intertidal zone. Similar experiments have been conducted on competition among desert rodents.

Competition and the Niches of Small Rodents

One of the most ambitious and complete of the many field experiments ecologists have conducted on competition among rodents focused on desert rodents in the Chihuahuan Desert near Portal, Arizona. This experiment, conducted by James H. Brown and his students and colleagues (Munger and Brown 1981, Brown and Munger 1985), is exceptional in many ways. First, it was conducted at a large scale; the 20 ha study site includes 24 study plots each 50 m by 50 m (fig. 13.22). Second, the experimental trials have been well replicated, both in space and in time. Third, the project has been long term; it began in 1977 and is ongoing. These three characteristics combine to demonstrate subtle ecological relationships and phenomena that would not otherwise be apparent.

FIGURE 13.22 Aerial photo showing the placement of 24 study plots. each 50 m by 50 m, in the Chihuahuan Desert near Portal. Arizona (courtesy of J. H. Brown).

The rodent species living on the Chihuahuan Desert study site can be divided into groups based upon size and feeding habits. Most members of the species are granivores, rodents that feed chiefly on seeds. The large granivores consist of three species of kangaroo rats (fig. 13.23a) in the genus Dipodomys-- D. spectabilis, 120 g; D. ordi, 52 g; and D. merriami, 45 g. In addition, the study site is home to four species of small granivores (fig. 13.23b)--Perognathus penicillatus, 17 g; P flavus, 7 g; Peromyscus maniculatus, 24 g; Reithrodontomys megalotis, 11 g -- and two species of small insectivorous rodents-- Onychomys leucogaster, 39 g; and O. torridus, 29 g.

FIGURE 13.23 Two species of granivorous rodents living in the Chihuahuan Desert: (a) the kangaroo rat, Dipodomys spp., a large granivore; (b) a pocket mouse. Pergonathus sp., a small granivore.

In one experiment, Brown and his colleagues set out to determine whether large granivorous rodents (Dipodomys spp.) limit the abundance of small rodents on their Chihnahuan Desert study site. They also wanted to know whether the rodents might be competing for food. The researchers addressed their questions with a field experiment in which they enclosed 50 m by 50 m study plots with mouse-proof fences. The fences were constructed with a wire mesh with 0.64 cm openings, which were too small for any of the rodent species to crawl through. They also buried the fencing 0.2 m deep so the mice couldn't dig under it, and they topped the fences with aluminum flashing so the mice couldn't climb over it. This may sound like a lot of work, but to answer their questions, the researchers had to control the presence of rodents on the study plots.

The researchers next cut holes 6.5 cm in diameter in the sides of all the fences to allow all rodent species to move freely in and out of the study plots. With this arrangement in place, the rodents in the study plots were trapped live and marked once a month for 3 months. Following this initial monitoring period, the holes on four of eight study plots were reduced to 1.9 cm, small enough to exclude Dipodomys but large enough to allow free movement of small rodents. Brown and his colleagues refer to these fences with small holes as semipermeable membranes, since they allow the movement of small rodents but exclude Dipodomys, the large granivores in this system.

If Dipodomys competes with small rodents, how would you expect populations of small rodents to respond to its removal? The density of small rodent populations should increase, right? If food is the limiting resource, would you expect granivorous and insectivorous rodents to respond differently to Dipodomys removal? The researchers predicted that if competition among rodents is mainly for food, then small granivorous rodent populations would increase in response to Dipodomys removal, while insectivorous rodents would show little or no response.

The results of the experiment were consistent with the predictions. During the first 3 years of the experiment, small granivores were approximately 3.5 times more abundant on the Dipodomys removal plots compared to the control plots, while populations of small insectivorous rodents did net increase significantly (fig. 13.24).

FIGURE 13.24 Responses by small granivorous and insectivorous rodents to removal of large granivorous Dipodomys species (data from Heske, Brown, and Mistry 1994)

The results presented in figure 13.24 support the hypothesis that Dipodomys spp. competitively suppress populations of small granivores. But would they do so again in response to another experimental manipulation? We cannot be certain unless we repeat the experiment. That's just what Edward Heske, James H. Brown, and Shahroukh Mistry (1994) did. In 1988, they selected eight other fenced study plots that they had been monitoring since 1977, installed their semipermeable barriers on four of the plots, and removed Dipodomys from them. The result was an almost immediate increase in small granivore populations on the removal plots (fig. 13.25). By reproducing the major results of the first experiment, this second experiment greatly strengthens the case for competition between large and small granivores at this Chihuahuan Desert site.

FIGURE 13.25 Responses of small granivorous and insectivorous rodents to a second removal experiment, which was preceded by several years of study before initiating Dipodomys removal (data from Heske, Brown, and Mistry 1994).

Character Displacement

Because interspecific competition reduces the fitness of competing individuals, those individuals that compete less should have higher fitness than individuals that compete more. Because the degree of competition is assumed to depend upon the degree of niche overlap, interspecific competition has been predicted to lead to directional selection for reduced niche overlap. This process of evolution toward niche divergence in the face of competition is called character displacement.

The Galapagos finches Geospiza fortis, the medium ground finch, and G. fuliginosa, the small ground finch, provide one of the most convincing cases of character displacement. These two species occur apart from each other, that is, they are allopatric, on Daphne Major and Los Hermanos Island and occur together, that is, they are sympatric, on the island of Santa Cruz (fig. 13.26). Where the two species are allopatric, they have very similar beak sizes. However, where they are sympatric, the sizes of their beaks do not overlap. The allopatric G. fortis on Daphne Major have smaller beaks than those sympatric with G. fuliginosa on Santa Cruz, while the G. fuliginosa on Los Hermanos Island have beaks that are significantly larger than those sympatric with G. fortis on Santa Cruz. Since beak size correlates with diet in Galapagos finches, we can say that the sympatric populations of the two species on Santa Cruz have different feeding niches. Natural selection has apparently favored divergence in the feeding niches of these sympatric populations (Lack 1947, Schluter, Price, and Grant 1985, and Grant 1986).

FIGURE 13.26 Evidence for character displacement in beak size in populations of the Galapagos finches Geospiza fortis and G. fuliginosa (data from Grant 1986).

A few other studies have demonstrated similar patterns of character displacement among a variety of animal species, including Cnemidophorus lizards on islands off Baja California, Anolis lizards on Caribbean islands, and sticklebacks inhabiting small lakes around Vancouver Island, Canada. Character displacement has also been observed in laboratory populations of bean weevils. Many studies have provided preliminary data suggesting character displacement among populations but not establishing definitive evidence. Why is that? The main mason is that a definitive demonstration requires a great deal of evidence that is difficult to provide.

Mark Taper and Ted Case (1992) list six criteria that must be met to build a definitive case for character displacement:

1. Morphological differences between a pair of sympatric species (e.g., G. fortis and G. fuliginosa on Santa Cruz Island) are statistically greater than the differences between allopatric populations of the same species (G. fortis on Daphne Major and G. fuliginosa on Los Hermanos Island).

2. The observed differences between sympatric and allopatric populations have a genetic basis.

3. Differences between sympatric and allopatric populations must have evolved in place and they must not be due to the sympatric and allopatric populations having been derived from different founder populations already differing in the character under study (e.g., beak size).

4. Variation in the character (e.g., beak size) must have a known effect on use of resources (e.g., seed sizes).

5. There must be demonstrated competition for the resource under question (e.g., food) and competition must be directly correlated with similarity in the character (e.g., overlap in beak size).

6. Differences in the character cannot be explained by differences in the resources available to sympatric and allopatric populations (e.g., differences in the availability of seeds on one island versus another).

You can see how difficult it would be to satisfy all six of these criteria. It is fitting that one of the few studies that addresses all six criteria reasonably well is that of the Galapagos finches (Grant 1986, Taper and Case 1992), in the place where Darwin started the whole discussion.

Evidence for Competition in Nature

What have we learned since Darwin initiated our enduring discussion of the role of competition as an organizing force in nature? The study of competition has gone through several phases. There was an early theoretical phase, followed by work with laboratory models, which was in turn followed by intensive observation and experimentation in the field. These phases were followed by a period of vigorous questioning of the assumption that competition is an important force in nature. This questioning forced renewed attention to careful experimental design and stimulated a reanalysis of past research in order to weigh the existing evidence bearing on competition in nature.

Two of the first analyses were those of Thomas Schoener (1983) and Joseph Connell (1983), both of whom reviewed the evidence provided by field experiments. Schoener, who reviewed over 150 field experiments on interspecific competition, reported that competition was found in 90% of the studies and among 76% of the species. Connell, who reviewed 527 experiments on 215 species, found evidence of interspecific competition in about 40% of the experiments and about 50% of the species. Why is there such a difference in these results? One reason is that the researchers analyzed different groups of studies and used different criteria for including studies in their analyses. Despite these differences, the analyses indicate that competition is an important force in the lives of many species. However, we may still have a biased estimate of the frequency of competition in nature, since ecologists may focus their studies of competition on species that they expect are competing.

About a decade after the analyses by Schoener and Connell, Jessica Gurevitch and her colleagues (1992) analyzed competition studies using a different statistical approach. Their analysis of studies conducted from 1980 to 1989 indicated a large overall effect of competition but also considerable differences among organisms and experimental approaches. They detected small to medium effects of competition on primary producers and carnivores and large effects on some herbivores and stream arthropods. Their analysis also indicated medium effects on herbivorous marine mollusks and echinoderms but no significant effects on herbivorous terrestrial insects. These researchers also found that larger experiments of longer duration yield less variable results than experiments of smaller scale and shorter duration.

After almost two decades of criticism, reflection, and reanalysis, what can we say about the prevalence and importance of interspecific competition in nature? The evidence supports the conclusion that competition is a common and important force that contributes to the organization of nature. However, the evidence also indicates that competition is neither omnipresent nor omnipotent. What other forces besides competition may be responsible for the patterns of distribution and abundance that we observe in natural populations? We've already reviewed the influences of the physical environment (sections I, II, and III) and in the next two chapters

we consider two other forms of biotic interaction: exploitation (chapter 14) and mutualism (chapter 15). But first, let's consider in detail one of the most important tools in the ecologist's tool kit, the field experiment.

APPLICATIONS AND TOOLS: the design of field experiments

Ecologists have built the present understanding of ecological relationships by gathering information, which they extract from nature by applying the tools of science. What are the scientific "tools" of ecology? As in other areas of science, ecologists use a wide range of tools that can be classified as "hardware." This hardware ranges from commonplace items such as microscopes and binoculars to high-tech items like supercomputers and satellites equipped with specialized sensors for monitoring environmental conditions. As we have seen, however, not all the tools of science consist of hardware. Some of the most important are conceptual or procedural tools, such as hypotheses, mathematical models, and laboratory experiments. To these, ecology adds one of the most powerful procedural tools at the disposal of the field ecologist: the field experiment.

Field experiments have played a key role in the assessment of the importance of competitive interactions in nature, and it is important that all ecologists understand their uses and limitations. Joseph Connell (1974) and Nelson Hairston, Sr. (1989), two of the pioneers in the use of field experiments in ecology, outlined their proper design and execution. Connell points out that one of the most substantial differences between laboratory and field experiments is that in the laboratory setting, the investigator controls all important factors but one, the factor of interest. In contrast, in field experiments, all factors are allowed to vary naturally (the investigator generally has no choice) while the factor of interest is controlled, or manipulated, by the investigator.

Both laboratory and field experiments have played an important role in ecology, but it is the field experiment that provides the key to unlock the secrets of complex interactions in nature. Why is it that field experiments are more useful in this regard? Connell pointed out that compared to laboratory experiments, the results of field experiments can be more directly applied to understanding relationships in nature because "interactions with other organisms, and the natural variation in the abiotic environment, are included in the experiment." The best field experiments are those that are executed with the least disturbance to the natural community. The utility of field experiments, however, depends upon several design features.

A Testable Hypothesis

One of the most basic requirements of a properly designed field experiment is a testable hypothesis. Formulating one is not as easy as it might sound. Posing any interesting hypothesis requires a great deal of thought and a lot of preliminary information about the organisms and environment to be studied. Posing an interesting hypothesis that is also testable is even more difficult. Field experimentation requires that you see beyond the complexities of nature to an experimental design that will reveal the factors contributing to that complexity.

Knowledge of Initial Conditions

To test for change in response to experimental manipulation, you have to know what conditions were like before the manipulation. Departures from initial conditions indicate a response to the experimental treatments. For instance, in his experiments on competition between barnacles, Connell first estimated the initial population density of one of the species in all his study plots (see fig. 13.20). Brown and his colleagues were also careful to measure the population densities of all rodent species in their study plots several times before excluding large granivorous rodents from their experimental plots.

Controls

As in laboratory experiments, field experiments must include controls. Without controls it would be impossible to determine whether or not an experimental treatment has had an effect. Tansley created controls for his experiments on competition by growing each of his potential competitors by themselves in acidic and basic soils. What was the control for the experiment on competition among desert rodents? Brown's research team created controls by surrounding study plots with their mouse-proof fence but then cutting holes 6.5 cm in diameter in the fences to allow large granivorous rodents to move freely into and out of the plots.

The field ecologist must take special care in the design of controls. For example, why couldn't Brown's research team have just compared the population densities of small granivores on their large granivore removal plots to the population density of small granivores on the surrounding open desert? Why did they have to go through the trouble of completely fencing their control plots only to cut holes in the sides? The reason is that the fence itself and the activity associated with its construction may have had an effect on rodent densities. Without their carefully constructed controls, they could not have concluded that the increase in small granivore densities on their experimental plots was the result of removing large granivores.

Replication

Why must experiments be done more than once at one site? The reason is that the same experiment will yield somewhat different results if done at different times or in different plots at the same time. The bottom line is that ecological systems and environmental conditions are variable, both in time and space. Replication captures this variation in response. The question posed by the experimenter is whether an experimental effect is apparent despite variation. Ecologists use statistics to make such a judgment. Without replication, you would never know if the results could be repeated either in time or space.

What is considered acceptable study design has changed over the decades, reflecting increasing familiarity and concern for statistical analysis. In Tansley's experiments on how competition may restrict the distribution of Galium species to particular soils, replication was totally lacking. In Tansley's experiment each condition (soil type) was represented just one time. Connell's later experiments with barnacles included some replication, but it was still limited at each tidal level. However, since there was a great deal of consistency in response across tidal levels, we can accept that competition acts as a significant force limiting barnacle distributions within the intertidal zone. In contrast to these earlier experiments, the more recent experiments by Brown on competition among desert rodents were replicated sufficiently for statistical analysis and repeated a second rime.

The reviews by Connell and Hairston provide a guide to field experimentation as it has been conducted in the past few decades. However, as we shall see in section VI, the need for experimentation at large scales is forcing ecologists to further expand their concept of experimental design.

SUMMARY CONCEPTS

Competition, interactions between individuals that reduce their fitness, is generally divided into intraspecific competition, competition between individuals of the same species, and interspecific competition, competition between individuals of different species. Competition can take the form of interference competition, direct aggressive interactions, or resource competition in which individuals compete through their dependence on the same limiting resources.

Studies of intraspecific competition provide evidence for resource limitation. Experiments with herbaceous plants show that soil nutrients may limit plant growth rates and plant weight and that competition for nutrients increases with plant population density. Plants reflect their competition for resources, including water, light, and nutrients, through the process of self-thinning. Resource competition among leafhoppers also varies with population density and is reflected in reduced survivorship, smaller size, and increased developmental time at higher population densities. Experiments with terrestrial isopods show that even where there is no food shortage, intraspecific competition through interference may be substantial.

The niche reflects the environmental requirements of species. The niche concept was developed early in the history of ecology and has had a prominent place in the study of interspecific competition because of the competitive exclusion principle: two species with identical niches cannot coexist indefinitely. Hutchinson added the concepts of the fundamental niche, the physical conditions under which a species might live in the absence of other species, and the realized niche, the more restricted conditions under which a species actually lives as the result of interactions with other species. While a species' niche is theoretically defined by a very large number of biotic and abiotic factors, Hutchinson's n-dimensional hypervolume, the most important attributes of the niche of most species, can often be summarized by a few variables. For instance, the niches of Galapagos finches are largely determined by their feeding requirements, while the niche of a salt marsh grass can be defined by tidal levels.

Mathematical and laboratory models provide a theoretical foundation for studying competitive interactions in nature. Lotka and Volterra independently expanded the logistic model of population growth to represent interspecific competition. In the Lotka-Volterra competition model, the growth rate of a species depends both upon numbers of conspecifics and numbers of the competing species. In this model, the effect of one species upon another is summarized by competition coefficients. In general, the Lotka-Volterra competition model predicts coexistence of species when interspecific competition is less intense than intraspecific competition. Competitive exclusion in laboratory experiments suggests the potential for competitive exclusion in nature. Even in the laboratory, however, organisms yield results that are much less predictable than the predictions of the Lotka-Volterra competition equations.

Competition can have significant ecological and evolutionary influences on the niches of species. Field experiments involving organisms from herbaceous plants to desert rodents have demonstrated that competition can restrict the niches of species to a narrower set of conditions than they would otherwise occupy in the absence of competition. Theoretically, natural selection may lead to divergence in the niches of competing species, a situation called character displacement. However, stringent requirements for a definitive demonstration have limited the number of documented cases of character displacement in nature. After many decades of

theoretical and experimental work on competition, we can conclude that competition is a common and strong force operating in nature, but not always and not everywhere.

The field experiment is one of the most powerful and important tools at the disposal of the field ecologist. However, the validity of field experiments depends upon several design features, including (1) a testable hypothesis, (2) knowedge of initial conditions, (3) controls, and (4) replication.

REVIEW QUESTIONS

1. Design a greenhouse (glasshouse) experiment to test for intraspecific competition within a population of herbivorous plants. Specify the species of plant, the volume (or size of pot) and source of soil, the potentially limiting resource you will focus on (e.g., Tilman and Cowan [1989] studied competition for nitrogen) and how you will manipulate it, and the measures of plant performance you will make.

2. How can the results of greenhouse experiments on competition help us understand the importance of competition among natural populations? How can a researcher enhance the correspondence of results between greenhouse experiments and the field situation?

3. Explain how self-thinning in field populations of plants can be used to support the hypothesis that intraspecific competition is a common occurrence among natural plant populations.

4. Researchers have characterized the niches of Galapagos finches by beak size (which correlates with diet) and the niches of salt marsh grasses by position in the intertidal zone. How would you characterize the niches of sympatric canid species such as red fox, coyote, and wolf in North America? Or felids such as ocelots, pumas, and jaguars in South America? What characteristics or environmental features do you think would be useful for representing the niches of desert plants? Or the plants in temperate forest or prairie?

5. Explain why species that overlap a great deal in their fundamental niches have a high probability of competing. Now explain why species that overlap a great deal in their realized niches and live in the same area probably do not compete significantly.

6. Draw the four possible ways in which Lotka's (1932) isoclines of zero growth (see fig. 13.14) can be oriented with respect to each other. Label the axes and the points where the isoclines intersect the horizontal and vertical axes. Explain how each situation rep- resented by the graphs leads to either competitive exclusion of one species or the other or to stable or unstable coexistence.

7. How was the amount of food that Gause (1934) provided in his experiment on competition among paramecia related to carrying capacity? In Gause's experiments on competition, P. aurelia excluded P. caudatum faster when he provided half the amount of food than when he doubled the amount of food. Explain.

8. In his experiments on competition between T. confusum and T. castaneum, Park (1954) found that one species usually excluded the other species but that the outcome depended upon physical conditions. In which circumstances did T. confusum have the competitive advantage? In which circumstances did T. castaneum have the competitive advantage? Could Park predict the outcomes of these experiments with complete certainty? What does this suggest about competition in nature? Discuss how mathematical theory, laboratory models, and field experiments have contributed to our understanding of the ecology of competition. List the advantages and disadvantages of each approach.

10. One of the conclusions that seems justified in light of several decades of studies of interspecific competition is that competition is a common and strong force operating in nature, but not always and not everywhere. List the environmental circumstances in which you think intraspecific and interspecific competition would be most likely to occur in nature. In what circumstances do you think competition is least likely to occur? How would you go about testing your ideas?

SUGGESTED READINGS

Lonsdale, W. M. 1990. The self-thinning rule: dead or alive? Ecology 71:1373-88.

Wellar, D. E. 1987. A reevaluation of the -3/2 power rule of plant self-thinning. Ecological Monographs 57:23-43.

Weller, D. E. 1989. The interspecific size-density relationship among crowded plant stands and its implications for the -3/2 power rule of self-thinning. American Naturalist 133:20-41.

Westoby, M. 1984. The self-thinning rule. Advances in Ecological Research 14:167-255.

This series of papers reviews the development and refinement of the self-thinning rule--an important principle of plant ecology with implications to the ecology of competition generally.

Hutchinson, G. E. 1957. Concluding remarks. Cold Spring Symposia on Quantitative Biology 22:415-27.

This is one of the most influential papers ever written on the subject of the niche by one of the most influential ecologists of the twentieth century.

Connell, J. H. 1961. The influence of interspecific competition and other factors on the distribution of the barnacle Chthamalus stellatus. Ecology 42:710-23.

This is one of the most enduring classics in the area of competition ecology and one of the early modem field experiments concerning interspecific competition.

Heske, E. J., J. H. Brown, and S. Mistry. 1994. Long-term experimental study of a Chihuahuan Desert rodent community: 13 years of competition. Ecology 75:438-45.

This paper reports on the first 13 years of one of the most ambitious experiments on interspecific competition among terrestrial animals. Destined to become a classic.

Schoener, T. W. 1982. The controversy over interspecific competition. American Scientist 70:586-95.

Wiens, J. A. 1977. On competition and variable environments. American Scientist 65:590-97.

These two papers represent two sides of a historical debate regarding the importance of competition in nature. This debate had a major influence on the course of ecology during the 1970s and 1980s.

Connell, J. H..!983: On the prevalence and relative importance of interspecific competition: evidence from field experiments. American Naturalist 122:661-96.

Goldberg, D. E. and A. M. Barton. 1992. Patterns and consequences of interspecific competition in natural communities: a review of field experiments with plants. American Naturalist 139:771-801.

Gurevitch, J., L. L Morrow, A. Wallace, and J. S. Walsh. 1992. A meta-analysis of competition in field experiments. American Naturalist 140:539-72.

Schoener, T W. 1983. Field experiments on interspecific competition. American Naturalist 122:240-85.

These four papers trace the history of a controversy concerning the importance of competition in nature. The analyses and the thoughtful presentations by all authors provide insights into some of the uncertainties and means of resolving controversy in ecology.

Grant, P. R. 1994. Ecological character displacement. Science 266:746-47.

Schluter, D., T D. Price, and P. R. Grant. 1985. Ecological character displacement in Darwin's finches. Science 227:1056-59.

Schluter, D. 1994. Experimental evidence that competition promotes divergence in adaptive radiation. Science 266:798-801.

These papers provide some of the best documented examples and concise reviews of the topic of character displacement.

ON THE NET

Visit our website at http://www.mhhe.com/ecology for links

to the following topics:

Animal Population Ecology

Competition

Field Methods for Studies of Populations

Field Methods for Studies of Ecosystems

Coevolution